├── cloud_function
    ├── requirements.txt
    ├── create_cloud_function.sh
    └── main.py
├── readme.md
└── notebooks
    ├── update_timestamps.ipynb
    ├── txn_and_features_gen.ipynb
    └── Validating_Online_Features_While_Detecting_Fraud.ipynb


/cloud_function/requirements.txt:
--------------------------------------------------------------------------------
1 | pandas
2 | pandas-gbq


--------------------------------------------------------------------------------
/cloud_function/create_cloud_function.sh:
--------------------------------------------------------------------------------
 1 | gcloud functions deploy feast-update-timestamps \
 2 | --entry-point main \
 3 | --runtime python37 \
 4 | --trigger-resource feature-timestamp-schedule \
 5 | --trigger-event google.pubsub.topic.publish \
 6 | --timeout 540s
 7 | 
 8 | gcloud scheduler jobs create pubsub feast-update-timestamp-job \
 9 | --schedule "0 22 * * *" \
10 | --topic feature-timestamp-schedule \
11 | --message-body "."


--------------------------------------------------------------------------------
/readme.md:
--------------------------------------------------------------------------------
 1 | # Feast Fraud Tutorial
 2 | 
 3 | This repo contains resources backing the `notebooks`directory containing notebooks used to generate the raw data in BigQuery.
 4 | `cloud_functions` contains the source code for a cloud function that updates timestamps of tables in BigQuery.
 5 | 
 6 | ### Prerequisites
 7 | 
 8 | To successfully run the primary tutorial in `notebooks` directory, `Fraud Detection Tutorial`, you 
 9 | must meet the following requirements:
10 |  * A Google Cloud Platform (GCP) account
11 |  * Ability to create, modify or delete GCP artifacts:
12 |    * Project ID
13 |    * Bucket name and location on GCP
14 |    * BigQuery Dataset name
15 | 
16 | 
17 | 
18 | 


--------------------------------------------------------------------------------
/cloud_function/main.py:
--------------------------------------------------------------------------------
 1 | import pandas as pd
 2 | from datetime import datetime, timedelta
 3 | 
 4 | def update_transactions():
 5 |     sql = """
 6 |         SELECT *
 7 |         FROM `feast-oss.fraud_tutorial.transactions`
 8 |     """
 9 |     transactions = pd.read_gbq(sql, dialect='standard')
10 |     latest_time = transactions['timestamp'].max()
11 |     datediff = datetime.now() - latest_time.replace(tzinfo=None)
12 |     transactions['timestamp'] = transactions['timestamp'] + datediff
13 |     transactions.to_gbq(destination_table="fraud_tutorial.transactions", project_id="feast-oss", if_exists='replace')
14 | 
15 | def update_user_features():
16 |     sql = """
17 |         SELECT *
18 |         FROM `feast-oss.fraud_tutorial.user_account_features`
19 |     """
20 |     user_features = pd.read_gbq(sql, dialect='standard')
21 |     user_features['feature_timestamp'] = datetime.now() - timedelta(days=7)
22 |     user_features.to_gbq(destination_table="fraud_tutorial.user_account_features", project_id="feast-oss", if_exists='replace')
23 | 
24 | def update_user_fraud_features():
25 |     sql = """
26 |         SELECT *
27 |         FROM `feast-oss.fraud_tutorial.user_has_fraudulent_transactions`
28 |     """
29 |     user_has_fraud = pd.read_gbq(sql, dialect='standard')
30 |     latest_time = user_has_fraud['feature_timestamp'].max()
31 |     datediff = datetime.now() - latest_time.replace(tzinfo=None)
32 |     user_has_fraud['feature_timestamp'] = user_has_fraud['feature_timestamp'] + datediff
33 |     user_has_fraud.to_gbq(destination_table="fraud_tutorial.user_has_fraudulent_transactions", project_id="feast-oss", if_exists='replace')
34 | 
35 | def main(data, context):
36 |     update_transactions()
37 |     update_user_features()
38 |     update_user_fraud_features()
39 | 
40 | main(1, 1)


--------------------------------------------------------------------------------
/notebooks/update_timestamps.ipynb:
--------------------------------------------------------------------------------
  1 | {
  2 |  "metadata": {
  3 |   "language_info": {
  4 |    "codemirror_mode": {
  5 |     "name": "ipython",
  6 |     "version": 3
  7 |    },
  8 |    "file_extension": ".py",
  9 |    "mimetype": "text/x-python",
 10 |    "name": "python",
 11 |    "nbconvert_exporter": "python",
 12 |    "pygments_lexer": "ipython3",
 13 |    "version": "3.7.9"
 14 |   },
 15 |   "orig_nbformat": 2,
 16 |   "kernelspec": {
 17 |    "name": "python3",
 18 |    "display_name": "Python 3.7.9 64-bit"
 19 |   },
 20 |   "interpreter": {
 21 |    "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
 22 |   }
 23 |  },
 24 |  "nbformat": 4,
 25 |  "nbformat_minor": 2,
 26 |  "cells": [
 27 |   {
 28 |    "cell_type": "code",
 29 |    "execution_count": 1,
 30 |    "metadata": {},
 31 |    "outputs": [
 32 |     {
 33 |      "output_type": "stream",
 34 |      "name": "stderr",
 35 |      "text": [
 36 |       "Downloading: 100%|██████████| 100000/100000 [00:05<00:00, 19643.91rows/s]\n"
 37 |      ]
 38 |     }
 39 |    ],
 40 |    "source": [
 41 |     "import pandas as pd\n",
 42 |     "\n",
 43 |     "sql = \"\"\"\n",
 44 |     "    SELECT *\n",
 45 |     "    FROM `feast-oss.fraud_tutorial.transactions`\n",
 46 |     "\"\"\"\n",
 47 |     "\n",
 48 |     "# Run a Standard SQL query using the environment's default project\n",
 49 |     "transactions = pd.read_gbq(sql, dialect='standard')\n",
 50 |     "\n",
 51 |     "latest_time = transactions['timestamp'].max()"
 52 |    ]
 53 |   },
 54 |   {
 55 |    "cell_type": "code",
 56 |    "execution_count": 2,
 57 |    "metadata": {},
 58 |    "outputs": [
 59 |     {
 60 |      "output_type": "execute_result",
 61 |      "data": {
 62 |       "text/plain": [
 63 |        "      src_account   amount dest_account  is_fraud  \\\n",
 64 |        "0          0001mg  3012.44       ydnwlr         0   \n",
 65 |        "1          0001mg  4431.82       oijv7z         0   \n",
 66 |        "2          0001mg  3037.60       a6mrvu         0   \n",
 67 |        "3          0001mg  6322.63       bmihen         0   \n",
 68 |        "4          0001mg  9981.82       tk53lu         0   \n",
 69 |        "...           ...      ...          ...       ...   \n",
 70 |        "99995      zyvtf8  3609.00       u5s54p         1   \n",
 71 |        "99996      zz0sgh  6060.71       c97pdy         1   \n",
 72 |        "99997      zz0sgh  5543.38       dt60g4         1   \n",
 73 |        "99998      zzrx9o  5031.12       9vo8j7         1   \n",
 74 |        "99999      zzx65l  9031.58       p6w6un         1   \n",
 75 |        "\n",
 76 |        "                             timestamp  \n",
 77 |        "0     2021-06-16 12:52:25.074517+00:00  \n",
 78 |        "1     2021-06-13 11:47:21.535700+00:00  \n",
 79 |        "2     2021-06-11 20:51:21.873945+00:00  \n",
 80 |        "3     2021-06-11 13:46:35.364700+00:00  \n",
 81 |        "4     2021-06-08 23:31:54.140277+00:00  \n",
 82 |        "...                                ...  \n",
 83 |        "99995 2021-06-12 21:09:53.775954+00:00  \n",
 84 |        "99996 2021-06-04 11:50:31.591834+00:00  \n",
 85 |        "99997 2021-06-03 21:48:26.560339+00:00  \n",
 86 |        "99998 2021-06-14 12:00:42.439961+00:00  \n",
 87 |        "99999 2021-06-11 13:13:24.071963+00:00  \n",
 88 |        "\n",
 89 |        "[100000 rows x 5 columns]"
 90 |       ],
 91 |       "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>src_account</th>\n      <th>amount</th>\n      <th>dest_account</th>\n      <th>is_fraud</th>\n      <th>timestamp</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>0001mg</td>\n      <td>3012.44</td>\n      <td>ydnwlr</td>\n      <td>0</td>\n      <td>2021-06-16 12:52:25.074517+00:00</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>0001mg</td>\n      <td>4431.82</td>\n      <td>oijv7z</td>\n      <td>0</td>\n      <td>2021-06-13 11:47:21.535700+00:00</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>0001mg</td>\n      <td>3037.60</td>\n      <td>a6mrvu</td>\n      <td>0</td>\n      <td>2021-06-11 20:51:21.873945+00:00</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>0001mg</td>\n      <td>6322.63</td>\n      <td>bmihen</td>\n      <td>0</td>\n      <td>2021-06-11 13:46:35.364700+00:00</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>0001mg</td>\n      <td>9981.82</td>\n      <td>tk53lu</td>\n      <td>0</td>\n      <td>2021-06-08 23:31:54.140277+00:00</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>99995</th>\n      <td>zyvtf8</td>\n      <td>3609.00</td>\n      <td>u5s54p</td>\n      <td>1</td>\n      <td>2021-06-12 21:09:53.775954+00:00</td>\n    </tr>\n    <tr>\n      <th>99996</th>\n      <td>zz0sgh</td>\n      <td>6060.71</td>\n      <td>c97pdy</td>\n      <td>1</td>\n      <td>2021-06-04 11:50:31.591834+00:00</td>\n    </tr>\n    <tr>\n      <th>99997</th>\n      <td>zz0sgh</td>\n      <td>5543.38</td>\n      <td>dt60g4</td>\n      <td>1</td>\n      <td>2021-06-03 21:48:26.560339+00:00</td>\n    </tr>\n    <tr>\n      <th>99998</th>\n      <td>zzrx9o</td>\n      <td>5031.12</td>\n      <td>9vo8j7</td>\n      <td>1</td>\n      <td>2021-06-14 12:00:42.439961+00:00</td>\n    </tr>\n    <tr>\n      <th>99999</th>\n      <td>zzx65l</td>\n      <td>9031.58</td>\n      <td>p6w6un</td>\n      <td>1</td>\n      <td>2021-06-11 13:13:24.071963+00:00</td>\n    </tr>\n  </tbody>\n</table>\n<p>100000 rows × 5 columns</p>\n</div>"
 92 |      },
 93 |      "metadata": {},
 94 |      "execution_count": 2
 95 |     }
 96 |    ],
 97 |    "source": [
 98 |     "from datetime import datetime\n",
 99 |     "datediff = datetime.now() - latest_time.replace(tzinfo=None)\n",
100 |     "\n",
101 |     "transactions['timestamp'] = transactions['timestamp'] + datediff\n"
102 |    ]
103 |   },
104 |   {
105 |    "cell_type": "code",
106 |    "execution_count": 3,
107 |    "metadata": {},
108 |    "outputs": [
109 |     {
110 |      "output_type": "stream",
111 |      "name": "stderr",
112 |      "text": [
113 |       "1it [00:11, 11.66s/it]\n"
114 |      ]
115 |     }
116 |    ],
117 |    "source": [
118 |     "transactions.to_gbq(destination_table=\"fraud_tutorial.transactions\", project_id=\"feast-oss\", if_exists='replace')"
119 |    ]
120 |   },
121 |   {
122 |    "cell_type": "code",
123 |    "execution_count": 4,
124 |    "metadata": {},
125 |    "outputs": [
126 |     {
127 |      "output_type": "stream",
128 |      "name": "stderr",
129 |      "text": [
130 |       "Downloading: 100%|██████████| 9944/9944 [00:00<00:00, 19073.20rows/s]\n"
131 |      ]
132 |     },
133 |     {
134 |      "output_type": "execute_result",
135 |      "data": {
136 |       "text/plain": [
137 |        "     user_id  credit_score  account_age_days  user_has_2fa_installed  \\\n",
138 |        "0     41sozr           512               700                       0   \n",
139 |        "1     h8nr8u           512               157                       0   \n",
140 |        "2     shid6v           512               509                       0   \n",
141 |        "3     rbcoqw           512               742                       0   \n",
142 |        "4     hew545           512               327                       0   \n",
143 |        "...      ...           ...               ...                     ...   \n",
144 |        "9939  nsgtkp           767               891                       1   \n",
145 |        "9940  4dlidj           767               855                       1   \n",
146 |        "9941  1z87hk           767               271                       1   \n",
147 |        "9942  ffqerm           767               829                       1   \n",
148 |        "9943  elz674           767               783                       1   \n",
149 |        "\n",
150 |        "              feature_timestamp  \n",
151 |        "0    2021-06-09 19:13:46.199693  \n",
152 |        "1    2021-06-09 19:13:46.199693  \n",
153 |        "2    2021-06-09 19:13:46.199693  \n",
154 |        "3    2021-06-09 19:13:46.199693  \n",
155 |        "4    2021-06-09 19:13:46.199693  \n",
156 |        "...                         ...  \n",
157 |        "9939 2021-06-09 19:13:46.199693  \n",
158 |        "9940 2021-06-09 19:13:46.199693  \n",
159 |        "9941 2021-06-09 19:13:46.199693  \n",
160 |        "9942 2021-06-09 19:13:46.199693  \n",
161 |        "9943 2021-06-09 19:13:46.199693  \n",
162 |        "\n",
163 |        "[9944 rows x 5 columns]"
164 |       ],
165 |       "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>user_id</th>\n      <th>credit_score</th>\n      <th>account_age_days</th>\n      <th>user_has_2fa_installed</th>\n      <th>feature_timestamp</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>41sozr</td>\n      <td>512</td>\n      <td>700</td>\n      <td>0</td>\n      <td>2021-06-09 19:13:46.199693</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>h8nr8u</td>\n      <td>512</td>\n      <td>157</td>\n      <td>0</td>\n      <td>2021-06-09 19:13:46.199693</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>shid6v</td>\n      <td>512</td>\n      <td>509</td>\n      <td>0</td>\n      <td>2021-06-09 19:13:46.199693</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>rbcoqw</td>\n      <td>512</td>\n      <td>742</td>\n      <td>0</td>\n      <td>2021-06-09 19:13:46.199693</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>hew545</td>\n      <td>512</td>\n      <td>327</td>\n      <td>0</td>\n      <td>2021-06-09 19:13:46.199693</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>9939</th>\n      <td>nsgtkp</td>\n      <td>767</td>\n      <td>891</td>\n      <td>1</td>\n      <td>2021-06-09 19:13:46.199693</td>\n    </tr>\n    <tr>\n      <th>9940</th>\n      <td>4dlidj</td>\n      <td>767</td>\n      <td>855</td>\n      <td>1</td>\n      <td>2021-06-09 19:13:46.199693</td>\n    </tr>\n    <tr>\n      <th>9941</th>\n      <td>1z87hk</td>\n      <td>767</td>\n      <td>271</td>\n      <td>1</td>\n      <td>2021-06-09 19:13:46.199693</td>\n    </tr>\n    <tr>\n      <th>9942</th>\n      <td>ffqerm</td>\n      <td>767</td>\n      <td>829</td>\n      <td>1</td>\n      <td>2021-06-09 19:13:46.199693</td>\n    </tr>\n    <tr>\n      <th>9943</th>\n      <td>elz674</td>\n      <td>767</td>\n      <td>783</td>\n      <td>1</td>\n      <td>2021-06-09 19:13:46.199693</td>\n    </tr>\n  </tbody>\n</table>\n<p>9944 rows × 5 columns</p>\n</div>"
166 |      },
167 |      "metadata": {},
168 |      "execution_count": 4
169 |     }
170 |    ],
171 |    "source": [
172 |     "import pandas as pd\n",
173 |     "from datetime import datetime, timedelta\n",
174 |     "\n",
175 |     "sql = \"\"\"\n",
176 |     "    SELECT *\n",
177 |     "    FROM `feast-oss.fraud_tutorial.user_account_features`\n",
178 |     "\"\"\"\n",
179 |     "\n",
180 |     "user_features = pd.read_gbq(sql, dialect='standard')\n",
181 |     "\n",
182 |     "user_features['feature_timestamp'] = datetime.now() - timedelta(days=7)\n",
183 |     "\n",
184 |     "user_features"
185 |    ]
186 |   },
187 |   {
188 |    "cell_type": "code",
189 |    "execution_count": 5,
190 |    "metadata": {},
191 |    "outputs": [
192 |     {
193 |      "output_type": "stream",
194 |      "name": "stderr",
195 |      "text": [
196 |       "1it [00:02,  2.60s/it]\n"
197 |      ]
198 |     }
199 |    ],
200 |    "source": [
201 |     "user_features.to_gbq(destination_table=\"fraud_tutorial.user_account_features\", project_id=\"feast-oss\", if_exists='replace')"
202 |    ]
203 |   },
204 |   {
205 |    "cell_type": "code",
206 |    "execution_count": 7,
207 |    "metadata": {},
208 |    "outputs": [
209 |     {
210 |      "output_type": "stream",
211 |      "name": "stderr",
212 |      "text": [
213 |       "Downloading: 100%|██████████| 69608/69608 [00:02<00:00, 29059.20rows/s]\n"
214 |      ]
215 |     },
216 |     {
217 |      "output_type": "execute_result",
218 |      "data": {
219 |       "text/plain": [
220 |        "      user_id  user_has_fraudulent_transactions_7d  \\\n",
221 |        "0      0001mg                                  0.0   \n",
222 |        "1      00c8mc                                  0.0   \n",
223 |        "2      00gmwi                                  0.0   \n",
224 |        "3      00mbm9                                  0.0   \n",
225 |        "4      00wjqi                                  0.0   \n",
226 |        "...       ...                                  ...   \n",
227 |        "69603  54r2jp                                  1.0   \n",
228 |        "69604  phvjnv                                  1.0   \n",
229 |        "69605  vr9qpk                                  1.0   \n",
230 |        "69606  wija9d                                  1.0   \n",
231 |        "69607  yvkh8e                                  1.0   \n",
232 |        "\n",
233 |        "                     feature_timestamp  \n",
234 |        "0     2021-06-11 12:56:59.739937+00:00  \n",
235 |        "1     2021-06-11 12:56:59.739937+00:00  \n",
236 |        "2     2021-06-11 12:56:59.739937+00:00  \n",
237 |        "3     2021-06-11 12:56:59.739937+00:00  \n",
238 |        "4     2021-06-11 12:56:59.739937+00:00  \n",
239 |        "...                                ...  \n",
240 |        "69603 2021-06-17 12:56:59.739937+00:00  \n",
241 |        "69604 2021-06-17 12:56:59.739937+00:00  \n",
242 |        "69605 2021-06-17 12:56:59.739937+00:00  \n",
243 |        "69606 2021-06-17 12:56:59.739937+00:00  \n",
244 |        "69607 2021-06-17 12:56:59.739937+00:00  \n",
245 |        "\n",
246 |        "[69608 rows x 3 columns]"
247 |       ],
248 |       "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>user_id</th>\n      <th>user_has_fraudulent_transactions_7d</th>\n      <th>feature_timestamp</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>0001mg</td>\n      <td>0.0</td>\n      <td>2021-06-11 12:56:59.739937+00:00</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>00c8mc</td>\n      <td>0.0</td>\n      <td>2021-06-11 12:56:59.739937+00:00</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>00gmwi</td>\n      <td>0.0</td>\n      <td>2021-06-11 12:56:59.739937+00:00</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>00mbm9</td>\n      <td>0.0</td>\n      <td>2021-06-11 12:56:59.739937+00:00</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>00wjqi</td>\n      <td>0.0</td>\n      <td>2021-06-11 12:56:59.739937+00:00</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>69603</th>\n      <td>54r2jp</td>\n      <td>1.0</td>\n      <td>2021-06-17 12:56:59.739937+00:00</td>\n    </tr>\n    <tr>\n      <th>69604</th>\n      <td>phvjnv</td>\n      <td>1.0</td>\n      <td>2021-06-17 12:56:59.739937+00:00</td>\n    </tr>\n    <tr>\n      <th>69605</th>\n      <td>vr9qpk</td>\n      <td>1.0</td>\n      <td>2021-06-17 12:56:59.739937+00:00</td>\n    </tr>\n    <tr>\n      <th>69606</th>\n      <td>wija9d</td>\n      <td>1.0</td>\n      <td>2021-06-17 12:56:59.739937+00:00</td>\n    </tr>\n    <tr>\n      <th>69607</th>\n      <td>yvkh8e</td>\n      <td>1.0</td>\n      <td>2021-06-17 12:56:59.739937+00:00</td>\n    </tr>\n  </tbody>\n</table>\n<p>69608 rows × 3 columns</p>\n</div>"
249 |      },
250 |      "metadata": {},
251 |      "execution_count": 7
252 |     }
253 |    ],
254 |    "source": [
255 |     "import pandas as pd\n",
256 |     "\n",
257 |     "sql = \"\"\"\n",
258 |     "    SELECT *\n",
259 |     "    FROM `feast-oss.fraud_tutorial.user_has_fraudulent_transactions`\n",
260 |     "\"\"\"\n",
261 |     "\n",
262 |     "# Run a Standard SQL query using the environment's default project\n",
263 |     "user_has_fraud = pd.read_gbq(sql, dialect='standard')\n",
264 |     "\n",
265 |     "latest_time = user_has_fraud['feature_timestamp'].max()\n",
266 |     "\n",
267 |     "datediff = datetime.now() - latest_time.replace(tzinfo=None)\n",
268 |     "\n",
269 |     "user_has_fraud['feature_timestamp'] = user_has_fraud['feature_timestamp'] + datediff"
270 |    ]
271 |   },
272 |   {
273 |    "cell_type": "code",
274 |    "execution_count": 8,
275 |    "metadata": {},
276 |    "outputs": [
277 |     {
278 |      "output_type": "stream",
279 |      "name": "stderr",
280 |      "text": [
281 |       "1it [00:05,  5.72s/it]\n"
282 |      ]
283 |     }
284 |    ],
285 |    "source": [
286 |     "user_has_fraud.to_gbq(destination_table=\"fraud_tutorial.user_has_fraudulent_transactions\", project_id=\"feast-oss\", if_exists='replace')"
287 |    ]
288 |   },
289 |   {
290 |    "cell_type": "code",
291 |    "execution_count": null,
292 |    "metadata": {},
293 |    "outputs": [],
294 |    "source": []
295 |   }
296 |  ]
297 | }


--------------------------------------------------------------------------------
/notebooks/txn_and_features_gen.ipynb:
--------------------------------------------------------------------------------
   1 | {
   2 |  "cells": [
   3 |   {
   4 |    "cell_type": "code",
   5 |    "execution_count": 1,
   6 |    "metadata": {},
   7 |    "outputs": [],
   8 |    "source": [
   9 |     "import string\n",
  10 |     "import random"
  11 |    ]
  12 |   },
  13 |   {
  14 |    "cell_type": "code",
  15 |    "execution_count": 61,
  16 |    "metadata": {},
  17 |    "outputs": [],
  18 |    "source": [
  19 |     "from datetime import datetime, timedelta\n",
  20 |     "import numpy as np\n",
  21 |     "\n",
  22 |     "src_accounts = [user_ids[abs(int(np.random.normal(5000, 2500)))% 10000] for _ in range(100000)]\n",
  23 |     "amounts = [ round(np.random.uniform (100, 10000), 2) for _ in range(100000)]\n",
  24 |     "dest_accounts = [''.join(random.choices(string.digits + string.ascii_lowercase, k=6)) for _ in range(100000)]\n",
  25 |     "is_frauds = [np.random.binomial(1, 0.05) for _ in range(100000)]\n",
  26 |     "\n",
  27 |     "timestamps = []\n",
  28 |     "ts = datetime.now() - timedelta(days=14)\n",
  29 |     "for i in range(100000):\n",
  30 |     "    timestamps.append(ts)\n",
  31 |     "    ts += timedelta(seconds=np.random.uniform(0, 24))"
  32 |    ]
  33 |   },
  34 |   {
  35 |    "cell_type": "code",
  36 |    "execution_count": 67,
  37 |    "metadata": {},
  38 |    "outputs": [
  39 |     {
  40 |      "data": {
  41 |       "text/html": [
  42 |        "<div>\n",
  43 |        "<style scoped>\n",
  44 |        "    .dataframe tbody tr th:only-of-type {\n",
  45 |        "        vertical-align: middle;\n",
  46 |        "    }\n",
  47 |        "\n",
  48 |        "    .dataframe tbody tr th {\n",
  49 |        "        vertical-align: top;\n",
  50 |        "    }\n",
  51 |        "\n",
  52 |        "    .dataframe thead th {\n",
  53 |        "        text-align: right;\n",
  54 |        "    }\n",
  55 |        "</style>\n",
  56 |        "<table border=\"1\" class=\"dataframe\">\n",
  57 |        "  <thead>\n",
  58 |        "    <tr style=\"text-align: right;\">\n",
  59 |        "      <th></th>\n",
  60 |        "      <th>src_account</th>\n",
  61 |        "      <th>amount</th>\n",
  62 |        "      <th>dest_account</th>\n",
  63 |        "      <th>is_fraud</th>\n",
  64 |        "      <th>timestamp</th>\n",
  65 |        "    </tr>\n",
  66 |        "  </thead>\n",
  67 |        "  <tbody>\n",
  68 |        "    <tr>\n",
  69 |        "      <th>0</th>\n",
  70 |        "      <td>7859ge</td>\n",
  71 |        "      <td>7078.83</td>\n",
  72 |        "      <td>waqmx5</td>\n",
  73 |        "      <td>0</td>\n",
  74 |        "      <td>2021-05-21 08:10:12.039737</td>\n",
  75 |        "    </tr>\n",
  76 |        "    <tr>\n",
  77 |        "      <th>1</th>\n",
  78 |        "      <td>yzziue</td>\n",
  79 |        "      <td>7851.83</td>\n",
  80 |        "      <td>tgx086</td>\n",
  81 |        "      <td>0</td>\n",
  82 |        "      <td>2021-05-21 08:10:24.464622</td>\n",
  83 |        "    </tr>\n",
  84 |        "    <tr>\n",
  85 |        "      <th>2</th>\n",
  86 |        "      <td>bgf8nl</td>\n",
  87 |        "      <td>6016.44</td>\n",
  88 |        "      <td>q0ltxc</td>\n",
  89 |        "      <td>0</td>\n",
  90 |        "      <td>2021-05-21 08:10:36.208894</td>\n",
  91 |        "    </tr>\n",
  92 |        "    <tr>\n",
  93 |        "      <th>3</th>\n",
  94 |        "      <td>jiaxoq</td>\n",
  95 |        "      <td>2573.43</td>\n",
  96 |        "      <td>ct01il</td>\n",
  97 |        "      <td>0</td>\n",
  98 |        "      <td>2021-05-21 08:10:54.177049</td>\n",
  99 |        "    </tr>\n",
 100 |        "    <tr>\n",
 101 |        "      <th>4</th>\n",
 102 |        "      <td>u49qmt</td>\n",
 103 |        "      <td>6743.81</td>\n",
 104 |        "      <td>397mqf</td>\n",
 105 |        "      <td>0</td>\n",
 106 |        "      <td>2021-05-21 08:11:16.870868</td>\n",
 107 |        "    </tr>\n",
 108 |        "    <tr>\n",
 109 |        "      <th>...</th>\n",
 110 |        "      <td>...</td>\n",
 111 |        "      <td>...</td>\n",
 112 |        "      <td>...</td>\n",
 113 |        "      <td>...</td>\n",
 114 |        "      <td>...</td>\n",
 115 |        "    </tr>\n",
 116 |        "    <tr>\n",
 117 |        "      <th>99995</th>\n",
 118 |        "      <td>xqjsd3</td>\n",
 119 |        "      <td>6856.42</td>\n",
 120 |        "      <td>2z3w39</td>\n",
 121 |        "      <td>0</td>\n",
 122 |        "      <td>2021-06-04 05:21:23.834089</td>\n",
 123 |        "    </tr>\n",
 124 |        "    <tr>\n",
 125 |        "      <th>99996</th>\n",
 126 |        "      <td>1tyh8p</td>\n",
 127 |        "      <td>8527.89</td>\n",
 128 |        "      <td>h5jgwy</td>\n",
 129 |        "      <td>0</td>\n",
 130 |        "      <td>2021-06-04 05:21:37.946295</td>\n",
 131 |        "    </tr>\n",
 132 |        "    <tr>\n",
 133 |        "      <th>99997</th>\n",
 134 |        "      <td>mfj3xt</td>\n",
 135 |        "      <td>4651.57</td>\n",
 136 |        "      <td>xvjv67</td>\n",
 137 |        "      <td>0</td>\n",
 138 |        "      <td>2021-06-04 05:21:39.853131</td>\n",
 139 |        "    </tr>\n",
 140 |        "    <tr>\n",
 141 |        "      <th>99998</th>\n",
 142 |        "      <td>l0e31n</td>\n",
 143 |        "      <td>9771.14</td>\n",
 144 |        "      <td>savnzy</td>\n",
 145 |        "      <td>0</td>\n",
 146 |        "      <td>2021-06-04 05:22:03.112553</td>\n",
 147 |        "    </tr>\n",
 148 |        "    <tr>\n",
 149 |        "      <th>99999</th>\n",
 150 |        "      <td>782k16</td>\n",
 151 |        "      <td>9749.99</td>\n",
 152 |        "      <td>24bhqv</td>\n",
 153 |        "      <td>0</td>\n",
 154 |        "      <td>2021-06-04 05:22:08.142090</td>\n",
 155 |        "    </tr>\n",
 156 |        "  </tbody>\n",
 157 |        "</table>\n",
 158 |        "<p>100000 rows × 5 columns</p>\n",
 159 |        "</div>"
 160 |       ],
 161 |       "text/plain": [
 162 |        "      src_account   amount dest_account  is_fraud                  timestamp\n",
 163 |        "0          7859ge  7078.83       waqmx5         0 2021-05-21 08:10:12.039737\n",
 164 |        "1          yzziue  7851.83       tgx086         0 2021-05-21 08:10:24.464622\n",
 165 |        "2          bgf8nl  6016.44       q0ltxc         0 2021-05-21 08:10:36.208894\n",
 166 |        "3          jiaxoq  2573.43       ct01il         0 2021-05-21 08:10:54.177049\n",
 167 |        "4          u49qmt  6743.81       397mqf         0 2021-05-21 08:11:16.870868\n",
 168 |        "...           ...      ...          ...       ...                        ...\n",
 169 |        "99995      xqjsd3  6856.42       2z3w39         0 2021-06-04 05:21:23.834089\n",
 170 |        "99996      1tyh8p  8527.89       h5jgwy         0 2021-06-04 05:21:37.946295\n",
 171 |        "99997      mfj3xt  4651.57       xvjv67         0 2021-06-04 05:21:39.853131\n",
 172 |        "99998      l0e31n  9771.14       savnzy         0 2021-06-04 05:22:03.112553\n",
 173 |        "99999      782k16  9749.99       24bhqv         0 2021-06-04 05:22:08.142090\n",
 174 |        "\n",
 175 |        "[100000 rows x 5 columns]"
 176 |       ]
 177 |      },
 178 |      "execution_count": 67,
 179 |      "metadata": {},
 180 |      "output_type": "execute_result"
 181 |     }
 182 |    ],
 183 |    "source": [
 184 |     "transactions = pd.DataFrame.from_dict({\n",
 185 |     "    \"src_account\": src_accounts,\n",
 186 |     "    \"amount\": amounts,\n",
 187 |     "    \"dest_account\": dest_accounts,\n",
 188 |     "    \"is_fraud\": is_frauds,\n",
 189 |     "    \"timestamp\": timestamps\n",
 190 |     "})\n",
 191 |     "\n",
 192 |     "transactions = transactions.sort_values(by='timestamp', ascending=False)\n",
 193 |     "\n",
 194 |     "transactions"
 195 |    ]
 196 |   },
 197 |   {
 198 |    "cell_type": "code",
 199 |    "execution_count": 138,
 200 |    "metadata": {},
 201 |    "outputs": [
 202 |     {
 203 |      "name": "stderr",
 204 |      "output_type": "stream",
 205 |      "text": [
 206 |       "1it [00:18, 18.60s/it]\n"
 207 |      ]
 208 |     }
 209 |    ],
 210 |    "source": [
 211 |     "# For some reason, this loads into BQ sorted by src_account\n",
 212 |     "# Needed to sort in BQ after this\n",
 213 |     "transactions.to_gbq(destination_table=\"fraud_tutorial.transactions\", project_id=\"feast-oss\")"
 214 |    ]
 215 |   },
 216 |   {
 217 |    "cell_type": "code",
 218 |    "execution_count": 4,
 219 |    "metadata": {},
 220 |    "outputs": [
 221 |     {
 222 |      "name": "stderr",
 223 |      "output_type": "stream",
 224 |      "text": [
 225 |       "Downloading: 100%|██████████| 100000/100000 [00:04<00:00, 23842.20rows/s]\n"
 226 |      ]
 227 |     }
 228 |    ],
 229 |    "source": [
 230 |     "import pandas as pd\n",
 231 |     "\n",
 232 |     "sql = \"\"\"\n",
 233 |     "    SELECT *\n",
 234 |     "    FROM `feast-oss.fraud_tutorial.transactions`\n",
 235 |     "\"\"\"\n",
 236 |     "\n",
 237 |     "# Run a Standard SQL query using the environment's default project\n",
 238 |     "transactions = pd.read_gbq(sql, dialect='standard')"
 239 |    ]
 240 |   },
 241 |   {
 242 |    "cell_type": "code",
 243 |    "execution_count": 6,
 244 |    "metadata": {},
 245 |    "outputs": [
 246 |     {
 247 |      "data": {
 248 |       "text/html": [
 249 |        "<div>\n",
 250 |        "<style scoped>\n",
 251 |        "    .dataframe tbody tr th:only-of-type {\n",
 252 |        "        vertical-align: middle;\n",
 253 |        "    }\n",
 254 |        "\n",
 255 |        "    .dataframe tbody tr th {\n",
 256 |        "        vertical-align: top;\n",
 257 |        "    }\n",
 258 |        "\n",
 259 |        "    .dataframe thead th {\n",
 260 |        "        text-align: right;\n",
 261 |        "    }\n",
 262 |        "</style>\n",
 263 |        "<table border=\"1\" class=\"dataframe\">\n",
 264 |        "  <thead>\n",
 265 |        "    <tr style=\"text-align: right;\">\n",
 266 |        "      <th></th>\n",
 267 |        "      <th>src_account</th>\n",
 268 |        "      <th>amount</th>\n",
 269 |        "      <th>dest_account</th>\n",
 270 |        "      <th>is_fraud</th>\n",
 271 |        "      <th>timestamp</th>\n",
 272 |        "    </tr>\n",
 273 |        "  </thead>\n",
 274 |        "  <tbody>\n",
 275 |        "    <tr>\n",
 276 |        "      <th>0</th>\n",
 277 |        "      <td>782k16</td>\n",
 278 |        "      <td>9749.99</td>\n",
 279 |        "      <td>24bhqv</td>\n",
 280 |        "      <td>0</td>\n",
 281 |        "      <td>2021-06-04 05:22:08.142090+00:00</td>\n",
 282 |        "    </tr>\n",
 283 |        "    <tr>\n",
 284 |        "      <th>1</th>\n",
 285 |        "      <td>l0e31n</td>\n",
 286 |        "      <td>9771.14</td>\n",
 287 |        "      <td>savnzy</td>\n",
 288 |        "      <td>0</td>\n",
 289 |        "      <td>2021-06-04 05:22:03.112553+00:00</td>\n",
 290 |        "    </tr>\n",
 291 |        "    <tr>\n",
 292 |        "      <th>2</th>\n",
 293 |        "      <td>mfj3xt</td>\n",
 294 |        "      <td>4651.57</td>\n",
 295 |        "      <td>xvjv67</td>\n",
 296 |        "      <td>0</td>\n",
 297 |        "      <td>2021-06-04 05:21:39.853131+00:00</td>\n",
 298 |        "    </tr>\n",
 299 |        "    <tr>\n",
 300 |        "      <th>3</th>\n",
 301 |        "      <td>1tyh8p</td>\n",
 302 |        "      <td>8527.89</td>\n",
 303 |        "      <td>h5jgwy</td>\n",
 304 |        "      <td>0</td>\n",
 305 |        "      <td>2021-06-04 05:21:37.946295+00:00</td>\n",
 306 |        "    </tr>\n",
 307 |        "    <tr>\n",
 308 |        "      <th>4</th>\n",
 309 |        "      <td>xqjsd3</td>\n",
 310 |        "      <td>6856.42</td>\n",
 311 |        "      <td>2z3w39</td>\n",
 312 |        "      <td>0</td>\n",
 313 |        "      <td>2021-06-04 05:21:23.834089+00:00</td>\n",
 314 |        "    </tr>\n",
 315 |        "    <tr>\n",
 316 |        "      <th>...</th>\n",
 317 |        "      <td>...</td>\n",
 318 |        "      <td>...</td>\n",
 319 |        "      <td>...</td>\n",
 320 |        "      <td>...</td>\n",
 321 |        "      <td>...</td>\n",
 322 |        "    </tr>\n",
 323 |        "    <tr>\n",
 324 |        "      <th>99995</th>\n",
 325 |        "      <td>u49qmt</td>\n",
 326 |        "      <td>6743.81</td>\n",
 327 |        "      <td>397mqf</td>\n",
 328 |        "      <td>0</td>\n",
 329 |        "      <td>2021-05-21 08:11:16.870868+00:00</td>\n",
 330 |        "    </tr>\n",
 331 |        "    <tr>\n",
 332 |        "      <th>99996</th>\n",
 333 |        "      <td>jiaxoq</td>\n",
 334 |        "      <td>2573.43</td>\n",
 335 |        "      <td>ct01il</td>\n",
 336 |        "      <td>0</td>\n",
 337 |        "      <td>2021-05-21 08:10:54.177049+00:00</td>\n",
 338 |        "    </tr>\n",
 339 |        "    <tr>\n",
 340 |        "      <th>99997</th>\n",
 341 |        "      <td>bgf8nl</td>\n",
 342 |        "      <td>6016.44</td>\n",
 343 |        "      <td>q0ltxc</td>\n",
 344 |        "      <td>0</td>\n",
 345 |        "      <td>2021-05-21 08:10:36.208894+00:00</td>\n",
 346 |        "    </tr>\n",
 347 |        "    <tr>\n",
 348 |        "      <th>99998</th>\n",
 349 |        "      <td>yzziue</td>\n",
 350 |        "      <td>7851.83</td>\n",
 351 |        "      <td>tgx086</td>\n",
 352 |        "      <td>0</td>\n",
 353 |        "      <td>2021-05-21 08:10:24.464622+00:00</td>\n",
 354 |        "    </tr>\n",
 355 |        "    <tr>\n",
 356 |        "      <th>99999</th>\n",
 357 |        "      <td>7859ge</td>\n",
 358 |        "      <td>7078.83</td>\n",
 359 |        "      <td>waqmx5</td>\n",
 360 |        "      <td>0</td>\n",
 361 |        "      <td>2021-05-21 08:10:12.039737+00:00</td>\n",
 362 |        "    </tr>\n",
 363 |        "  </tbody>\n",
 364 |        "</table>\n",
 365 |        "<p>100000 rows × 5 columns</p>\n",
 366 |        "</div>"
 367 |       ],
 368 |       "text/plain": [
 369 |        "      src_account   amount dest_account  is_fraud  \\\n",
 370 |        "0          782k16  9749.99       24bhqv         0   \n",
 371 |        "1          l0e31n  9771.14       savnzy         0   \n",
 372 |        "2          mfj3xt  4651.57       xvjv67         0   \n",
 373 |        "3          1tyh8p  8527.89       h5jgwy         0   \n",
 374 |        "4          xqjsd3  6856.42       2z3w39         0   \n",
 375 |        "...           ...      ...          ...       ...   \n",
 376 |        "99995      u49qmt  6743.81       397mqf         0   \n",
 377 |        "99996      jiaxoq  2573.43       ct01il         0   \n",
 378 |        "99997      bgf8nl  6016.44       q0ltxc         0   \n",
 379 |        "99998      yzziue  7851.83       tgx086         0   \n",
 380 |        "99999      7859ge  7078.83       waqmx5         0   \n",
 381 |        "\n",
 382 |        "                             timestamp  \n",
 383 |        "0     2021-06-04 05:22:08.142090+00:00  \n",
 384 |        "1     2021-06-04 05:22:03.112553+00:00  \n",
 385 |        "2     2021-06-04 05:21:39.853131+00:00  \n",
 386 |        "3     2021-06-04 05:21:37.946295+00:00  \n",
 387 |        "4     2021-06-04 05:21:23.834089+00:00  \n",
 388 |        "...                                ...  \n",
 389 |        "99995 2021-05-21 08:11:16.870868+00:00  \n",
 390 |        "99996 2021-05-21 08:10:54.177049+00:00  \n",
 391 |        "99997 2021-05-21 08:10:36.208894+00:00  \n",
 392 |        "99998 2021-05-21 08:10:24.464622+00:00  \n",
 393 |        "99999 2021-05-21 08:10:12.039737+00:00  \n",
 394 |        "\n",
 395 |        "[100000 rows x 5 columns]"
 396 |       ]
 397 |      },
 398 |      "execution_count": 6,
 399 |      "metadata": {},
 400 |      "output_type": "execute_result"
 401 |     }
 402 |    ],
 403 |    "source": [
 404 |     "transactions"
 405 |    ]
 406 |   },
 407 |   {
 408 |    "cell_type": "code",
 409 |    "execution_count": 10,
 410 |    "metadata": {},
 411 |    "outputs": [],
 412 |    "source": [
 413 |     "from datetime import datetime, timedelta\n",
 414 |     "import pytz\n",
 415 |     "\n",
 416 |     "t = transactions\n",
 417 |     "\n",
 418 |     "aggregation_end = datetime.now(tz=pytz.UTC)\n",
 419 |     "\n",
 420 |     "# last week's transactions\n",
 421 |     "lwt = t[(t['timestamp'] < aggregation_end ) & (t['timestamp'] >= aggregation_end - timedelta(days=7))]\n",
 422 |     "\n",
 423 |     "# users with fraudulent transactions\n",
 424 |     "fu = lwt[lwt['is_fraud'] == 1]\\\n",
 425 |     "        .drop_duplicates(subset='src_account')[['src_account', 'is_fraud']]\\\n",
 426 |     "        .rename(columns={'is_fraud': 'user_has_fraudulent_transactions_7d'})        \n"
 427 |    ]
 428 |   },
 429 |   {
 430 |    "cell_type": "code",
 431 |    "execution_count": 11,
 432 |    "metadata": {},
 433 |    "outputs": [
 434 |     {
 435 |      "data": {
 436 |       "text/html": [
 437 |        "<div>\n",
 438 |        "<style scoped>\n",
 439 |        "    .dataframe tbody tr th:only-of-type {\n",
 440 |        "        vertical-align: middle;\n",
 441 |        "    }\n",
 442 |        "\n",
 443 |        "    .dataframe tbody tr th {\n",
 444 |        "        vertical-align: top;\n",
 445 |        "    }\n",
 446 |        "\n",
 447 |        "    .dataframe thead th {\n",
 448 |        "        text-align: right;\n",
 449 |        "    }\n",
 450 |        "</style>\n",
 451 |        "<table border=\"1\" class=\"dataframe\">\n",
 452 |        "  <thead>\n",
 453 |        "    <tr style=\"text-align: right;\">\n",
 454 |        "      <th></th>\n",
 455 |        "      <th>src_account</th>\n",
 456 |        "      <th>user_has_fraudulent_transactions_7d</th>\n",
 457 |        "      <th>feature_timestamp</th>\n",
 458 |        "    </tr>\n",
 459 |        "  </thead>\n",
 460 |        "  <tbody>\n",
 461 |        "    <tr>\n",
 462 |        "      <th>27</th>\n",
 463 |        "      <td>zqvbs4</td>\n",
 464 |        "      <td>1</td>\n",
 465 |        "      <td>2021-06-07 12:45:10.214026</td>\n",
 466 |        "    </tr>\n",
 467 |        "    <tr>\n",
 468 |        "      <th>37</th>\n",
 469 |        "      <td>a9l0te</td>\n",
 470 |        "      <td>1</td>\n",
 471 |        "      <td>2021-06-07 12:45:10.214026</td>\n",
 472 |        "    </tr>\n",
 473 |        "    <tr>\n",
 474 |        "      <th>62</th>\n",
 475 |        "      <td>z2lnqe</td>\n",
 476 |        "      <td>1</td>\n",
 477 |        "      <td>2021-06-07 12:45:10.214026</td>\n",
 478 |        "    </tr>\n",
 479 |        "    <tr>\n",
 480 |        "      <th>82</th>\n",
 481 |        "      <td>xv1ul5</td>\n",
 482 |        "      <td>1</td>\n",
 483 |        "      <td>2021-06-07 12:45:10.214026</td>\n",
 484 |        "    </tr>\n",
 485 |        "    <tr>\n",
 486 |        "      <th>112</th>\n",
 487 |        "      <td>6ua5v6</td>\n",
 488 |        "      <td>1</td>\n",
 489 |        "      <td>2021-06-07 12:45:10.214026</td>\n",
 490 |        "    </tr>\n",
 491 |        "    <tr>\n",
 492 |        "      <th>...</th>\n",
 493 |        "      <td>...</td>\n",
 494 |        "      <td>...</td>\n",
 495 |        "      <td>...</td>\n",
 496 |        "    </tr>\n",
 497 |        "    <tr>\n",
 498 |        "      <th>25188</th>\n",
 499 |        "      <td>kr123d</td>\n",
 500 |        "      <td>1</td>\n",
 501 |        "      <td>2021-06-07 12:45:10.214026</td>\n",
 502 |        "    </tr>\n",
 503 |        "    <tr>\n",
 504 |        "      <th>25243</th>\n",
 505 |        "      <td>y7dobz</td>\n",
 506 |        "      <td>1</td>\n",
 507 |        "      <td>2021-06-07 12:45:10.214026</td>\n",
 508 |        "    </tr>\n",
 509 |        "    <tr>\n",
 510 |        "      <th>25303</th>\n",
 511 |        "      <td>wija9d</td>\n",
 512 |        "      <td>1</td>\n",
 513 |        "      <td>2021-06-07 12:45:10.214026</td>\n",
 514 |        "    </tr>\n",
 515 |        "    <tr>\n",
 516 |        "      <th>25375</th>\n",
 517 |        "      <td>u269is</td>\n",
 518 |        "      <td>1</td>\n",
 519 |        "      <td>2021-06-07 12:45:10.214026</td>\n",
 520 |        "    </tr>\n",
 521 |        "    <tr>\n",
 522 |        "      <th>25378</th>\n",
 523 |        "      <td>8058vz</td>\n",
 524 |        "      <td>1</td>\n",
 525 |        "      <td>2021-06-07 12:45:10.214026</td>\n",
 526 |        "    </tr>\n",
 527 |        "  </tbody>\n",
 528 |        "</table>\n",
 529 |        "<p>1175 rows × 3 columns</p>\n",
 530 |        "</div>"
 531 |       ],
 532 |       "text/plain": [
 533 |        "      src_account  user_has_fraudulent_transactions_7d  \\\n",
 534 |        "27         zqvbs4                                    1   \n",
 535 |        "37         a9l0te                                    1   \n",
 536 |        "62         z2lnqe                                    1   \n",
 537 |        "82         xv1ul5                                    1   \n",
 538 |        "112        6ua5v6                                    1   \n",
 539 |        "...           ...                                  ...   \n",
 540 |        "25188      kr123d                                    1   \n",
 541 |        "25243      y7dobz                                    1   \n",
 542 |        "25303      wija9d                                    1   \n",
 543 |        "25375      u269is                                    1   \n",
 544 |        "25378      8058vz                                    1   \n",
 545 |        "\n",
 546 |        "               feature_timestamp  \n",
 547 |        "27    2021-06-07 12:45:10.214026  \n",
 548 |        "37    2021-06-07 12:45:10.214026  \n",
 549 |        "62    2021-06-07 12:45:10.214026  \n",
 550 |        "82    2021-06-07 12:45:10.214026  \n",
 551 |        "112   2021-06-07 12:45:10.214026  \n",
 552 |        "...                          ...  \n",
 553 |        "25188 2021-06-07 12:45:10.214026  \n",
 554 |        "25243 2021-06-07 12:45:10.214026  \n",
 555 |        "25303 2021-06-07 12:45:10.214026  \n",
 556 |        "25375 2021-06-07 12:45:10.214026  \n",
 557 |        "25378 2021-06-07 12:45:10.214026  \n",
 558 |        "\n",
 559 |        "[1175 rows x 3 columns]"
 560 |       ]
 561 |      },
 562 |      "execution_count": 11,
 563 |      "metadata": {},
 564 |      "output_type": "execute_result"
 565 |     }
 566 |    ],
 567 |    "source": [
 568 |     "fu"
 569 |    ]
 570 |   },
 571 |   {
 572 |    "cell_type": "code",
 573 |    "execution_count": 23,
 574 |    "metadata": {},
 575 |    "outputs": [
 576 |     {
 577 |      "data": {
 578 |       "text/html": [
 579 |        "<div>\n",
 580 |        "<style scoped>\n",
 581 |        "    .dataframe tbody tr th:only-of-type {\n",
 582 |        "        vertical-align: middle;\n",
 583 |        "    }\n",
 584 |        "\n",
 585 |        "    .dataframe tbody tr th {\n",
 586 |        "        vertical-align: top;\n",
 587 |        "    }\n",
 588 |        "\n",
 589 |        "    .dataframe thead th {\n",
 590 |        "        text-align: right;\n",
 591 |        "    }\n",
 592 |        "</style>\n",
 593 |        "<table border=\"1\" class=\"dataframe\">\n",
 594 |        "  <thead>\n",
 595 |        "    <tr style=\"text-align: right;\">\n",
 596 |        "      <th></th>\n",
 597 |        "      <th>user_id</th>\n",
 598 |        "      <th>user_has_fraudulent_transactions_7d</th>\n",
 599 |        "      <th>feature_timestamp</th>\n",
 600 |        "    </tr>\n",
 601 |        "  </thead>\n",
 602 |        "  <tbody>\n",
 603 |        "    <tr>\n",
 604 |        "      <th>0</th>\n",
 605 |        "      <td>782k16</td>\n",
 606 |        "      <td>0.0</td>\n",
 607 |        "      <td>2021-06-07 12:58:28.318652</td>\n",
 608 |        "    </tr>\n",
 609 |        "    <tr>\n",
 610 |        "      <th>1</th>\n",
 611 |        "      <td>l0e31n</td>\n",
 612 |        "      <td>0.0</td>\n",
 613 |        "      <td>2021-06-07 12:58:28.318652</td>\n",
 614 |        "    </tr>\n",
 615 |        "    <tr>\n",
 616 |        "      <th>2</th>\n",
 617 |        "      <td>mfj3xt</td>\n",
 618 |        "      <td>1.0</td>\n",
 619 |        "      <td>2021-06-07 12:58:28.318652</td>\n",
 620 |        "    </tr>\n",
 621 |        "    <tr>\n",
 622 |        "      <th>3</th>\n",
 623 |        "      <td>1tyh8p</td>\n",
 624 |        "      <td>0.0</td>\n",
 625 |        "      <td>2021-06-07 12:58:28.318652</td>\n",
 626 |        "    </tr>\n",
 627 |        "    <tr>\n",
 628 |        "      <th>4</th>\n",
 629 |        "      <td>xqjsd3</td>\n",
 630 |        "      <td>0.0</td>\n",
 631 |        "      <td>2021-06-07 12:58:28.318652</td>\n",
 632 |        "    </tr>\n",
 633 |        "    <tr>\n",
 634 |        "      <th>...</th>\n",
 635 |        "      <td>...</td>\n",
 636 |        "      <td>...</td>\n",
 637 |        "      <td>...</td>\n",
 638 |        "    </tr>\n",
 639 |        "    <tr>\n",
 640 |        "      <th>99995</th>\n",
 641 |        "      <td>u49qmt</td>\n",
 642 |        "      <td>0.0</td>\n",
 643 |        "      <td>2021-06-07 12:58:28.318652</td>\n",
 644 |        "    </tr>\n",
 645 |        "    <tr>\n",
 646 |        "      <th>99996</th>\n",
 647 |        "      <td>jiaxoq</td>\n",
 648 |        "      <td>0.0</td>\n",
 649 |        "      <td>2021-06-07 12:58:28.318652</td>\n",
 650 |        "    </tr>\n",
 651 |        "    <tr>\n",
 652 |        "      <th>99997</th>\n",
 653 |        "      <td>bgf8nl</td>\n",
 654 |        "      <td>0.0</td>\n",
 655 |        "      <td>2021-06-07 12:58:28.318652</td>\n",
 656 |        "    </tr>\n",
 657 |        "    <tr>\n",
 658 |        "      <th>99998</th>\n",
 659 |        "      <td>yzziue</td>\n",
 660 |        "      <td>0.0</td>\n",
 661 |        "      <td>2021-06-07 12:58:28.318652</td>\n",
 662 |        "    </tr>\n",
 663 |        "    <tr>\n",
 664 |        "      <th>99999</th>\n",
 665 |        "      <td>7859ge</td>\n",
 666 |        "      <td>0.0</td>\n",
 667 |        "      <td>2021-06-07 12:58:28.318652</td>\n",
 668 |        "    </tr>\n",
 669 |        "  </tbody>\n",
 670 |        "</table>\n",
 671 |        "<p>100000 rows × 3 columns</p>\n",
 672 |        "</div>"
 673 |       ],
 674 |       "text/plain": [
 675 |        "      user_id  user_has_fraudulent_transactions_7d          feature_timestamp\n",
 676 |        "0      782k16                                  0.0 2021-06-07 12:58:28.318652\n",
 677 |        "1      l0e31n                                  0.0 2021-06-07 12:58:28.318652\n",
 678 |        "2      mfj3xt                                  1.0 2021-06-07 12:58:28.318652\n",
 679 |        "3      1tyh8p                                  0.0 2021-06-07 12:58:28.318652\n",
 680 |        "4      xqjsd3                                  0.0 2021-06-07 12:58:28.318652\n",
 681 |        "...       ...                                  ...                        ...\n",
 682 |        "99995  u49qmt                                  0.0 2021-06-07 12:58:28.318652\n",
 683 |        "99996  jiaxoq                                  0.0 2021-06-07 12:58:28.318652\n",
 684 |        "99997  bgf8nl                                  0.0 2021-06-07 12:58:28.318652\n",
 685 |        "99998  yzziue                                  0.0 2021-06-07 12:58:28.318652\n",
 686 |        "99999  7859ge                                  0.0 2021-06-07 12:58:28.318652\n",
 687 |        "\n",
 688 |        "[100000 rows x 3 columns]"
 689 |       ]
 690 |      },
 691 |      "execution_count": 23,
 692 |      "metadata": {},
 693 |      "output_type": "execute_result"
 694 |     }
 695 |    ],
 696 |    "source": [
 697 |     "# feature table\n",
 698 |     "ff = pd.merge(t, fu, on='src_account',how='left')\\\n",
 699 |     "        .fillna(0)\\\n",
 700 |     "        [['src_account', 'user_has_fraudulent_transactions_7d']].rename(columns={\"src_account\": \"user_id\"})\n",
 701 |     "\n",
 702 |     "ff['feature_timestamp'] = datetime.now()\n",
 703 |     "\n",
 704 |     "ff"
 705 |    ]
 706 |   },
 707 |   {
 708 |    "cell_type": "code",
 709 |    "execution_count": 20,
 710 |    "metadata": {},
 711 |    "outputs": [
 712 |     {
 713 |      "output_type": "stream",
 714 |      "name": "stderr",
 715 |      "text": [
 716 |       "\n",
 717 |       "Downloading:   0%|          | 0/100000 [00:00<?, ?rows/s]\u001b[A\n",
 718 |       "Downloading: 100%|██████████| 100000/100000 [00:03<00:00, 25626.05rows/s]\n",
 719 |       "generating features as of 2021-06-03 16:42:14.250521+00:00\n",
 720 |       "\n",
 721 |       "1it [00:03,  3.78s/it]\n",
 722 |       "generating features as of 2021-06-04 16:42:14.250521+00:00\n",
 723 |       "\n",
 724 |       "1it [00:02,  2.77s/it]\n",
 725 |       "generating features as of 2021-06-05 16:42:14.250521+00:00\n",
 726 |       "\n",
 727 |       "1it [00:02,  2.72s/it]\n",
 728 |       "generating features as of 2021-06-06 16:42:14.250521+00:00\n",
 729 |       "\n",
 730 |       "1it [00:05,  5.13s/it]\n",
 731 |       "generating features as of 2021-06-07 16:42:14.250521+00:00\n",
 732 |       "\n",
 733 |       "1it [00:02,  2.27s/it]\n",
 734 |       "generating features as of 2021-06-08 16:42:14.250521+00:00\n",
 735 |       "\n",
 736 |       "1it [00:02,  2.61s/it]\n",
 737 |       "\n",
 738 |       "0it [00:00, ?it/s]\u001b[Agenerating features as of 2021-06-09 16:42:14.250521+00:00\n",
 739 |       "1it [00:05,  5.17s/it]\n"
 740 |      ]
 741 |     }
 742 |    ],
 743 |    "source": [
 744 |     "from datetime import datetime, timedelta\n",
 745 |     "import pytz\n",
 746 |     "\n",
 747 |     "sql = \"\"\"\n",
 748 |     "SELECT *\n",
 749 |     "FROM `feast-oss.fraud_tutorial.transactions`\n",
 750 |     "\"\"\"\n",
 751 |     "\n",
 752 |     "t = pd.read_gbq(sql, dialect='standard')\n",
 753 |     "\n",
 754 |     "def generate_fraud_features(aggregation_end):\n",
 755 |     "\n",
 756 |     "    user_ids = t['src_account'].drop_duplicates()\n",
 757 |     "\n",
 758 |     "    # last week's transactions\n",
 759 |     "    lwt = t[(t['timestamp'] < aggregation_end ) & (t['timestamp'] >= aggregation_end - timedelta(days=7))]\n",
 760 |     "\n",
 761 |     "    # users with fraudulent transactions\n",
 762 |     "    fu = lwt[lwt['is_fraud'] == 1]\\\n",
 763 |     "            .drop_duplicates(subset='src_account')[['src_account', 'is_fraud']]\\\n",
 764 |     "            .rename(columns={'is_fraud': 'user_has_fraudulent_transactions_7d'}) \\\n",
 765 |     "    # feature table\n",
 766 |     "    ff = pd.merge(user_ids, fu, on='src_account',how='left')\\\n",
 767 |     "            .fillna(0)\\\n",
 768 |     "            [['src_account', 'user_has_fraudulent_transactions_7d']].rename(columns={\"src_account\": \"user_id\"})\n",
 769 |     "\n",
 770 |     "    ff['feature_timestamp'] = aggregation_end\n",
 771 |     "    ff.to_gbq(destination_table=\"fraud_tutorial.user_has_fraudulent_transactions\", project_id=\"feast-oss\", if_exists='append')\n",
 772 |     "\n",
 773 |     "def backfill_features(start_date, interval, iters):\n",
 774 |     "    aggregation_end_date = start_date\n",
 775 |     "    for _ in range(iters):\n",
 776 |     "        print(f\"generating features as of {aggregation_end_date}\")\n",
 777 |     "        generate_fraud_features(aggregation_end=aggregation_end_date)\n",
 778 |     "        aggregation_end_date += interval\n",
 779 |     "    \n",
 780 |     "ff = backfill_features(\n",
 781 |     "    start_date=datetime.now(tz=pytz.UTC) - timedelta(days=7), \n",
 782 |     "    interval=timedelta(days=1), \n",
 783 |     "    iters=7\n",
 784 |     ")"
 785 |    ]
 786 |   },
 787 |   {
 788 |    "cell_type": "code",
 789 |    "execution_count": 24,
 790 |    "metadata": {},
 791 |    "outputs": [
 792 |     {
 793 |      "name": "stderr",
 794 |      "output_type": "stream",
 795 |      "text": [
 796 |       "1it [00:06,  6.93s/it]\n"
 797 |      ]
 798 |     }
 799 |    ],
 800 |    "source": [
 801 |     "ff.to_gbq(destination_table=\"fraud_tutorial.user_has_fraudulent_transactions\", project_id=\"feast-oss\", if_exists='replace')"
 802 |    ]
 803 |   },
 804 |   {
 805 |    "cell_type": "code",
 806 |    "execution_count": 25,
 807 |    "metadata": {},
 808 |    "outputs": [
 809 |     {
 810 |      "data": {
 811 |       "text/html": [
 812 |        "<div>\n",
 813 |        "<style scoped>\n",
 814 |        "    .dataframe tbody tr th:only-of-type {\n",
 815 |        "        vertical-align: middle;\n",
 816 |        "    }\n",
 817 |        "\n",
 818 |        "    .dataframe tbody tr th {\n",
 819 |        "        vertical-align: top;\n",
 820 |        "    }\n",
 821 |        "\n",
 822 |        "    .dataframe thead th {\n",
 823 |        "        text-align: right;\n",
 824 |        "    }\n",
 825 |        "</style>\n",
 826 |        "<table border=\"1\" class=\"dataframe\">\n",
 827 |        "  <thead>\n",
 828 |        "    <tr style=\"text-align: right;\">\n",
 829 |        "      <th></th>\n",
 830 |        "      <th>user_id</th>\n",
 831 |        "      <th>credit_score</th>\n",
 832 |        "      <th>account_age_days</th>\n",
 833 |        "      <th>user_has_2fa_installed</th>\n",
 834 |        "      <th>feature_timestamp</th>\n",
 835 |        "    </tr>\n",
 836 |        "  </thead>\n",
 837 |        "  <tbody>\n",
 838 |        "    <tr>\n",
 839 |        "      <th>0</th>\n",
 840 |        "      <td>782k16</td>\n",
 841 |        "      <td>626</td>\n",
 842 |        "      <td>799</td>\n",
 843 |        "      <td>1</td>\n",
 844 |        "      <td>2021-06-07 12:59:14.813413</td>\n",
 845 |        "    </tr>\n",
 846 |        "    <tr>\n",
 847 |        "      <th>1</th>\n",
 848 |        "      <td>l0e31n</td>\n",
 849 |        "      <td>648</td>\n",
 850 |        "      <td>889</td>\n",
 851 |        "      <td>1</td>\n",
 852 |        "      <td>2021-06-07 12:59:14.813418</td>\n",
 853 |        "    </tr>\n",
 854 |        "    <tr>\n",
 855 |        "      <th>2</th>\n",
 856 |        "      <td>mfj3xt</td>\n",
 857 |        "      <td>603</td>\n",
 858 |        "      <td>383</td>\n",
 859 |        "      <td>1</td>\n",
 860 |        "      <td>2021-06-07 12:59:14.813419</td>\n",
 861 |        "    </tr>\n",
 862 |        "    <tr>\n",
 863 |        "      <th>3</th>\n",
 864 |        "      <td>1tyh8p</td>\n",
 865 |        "      <td>808</td>\n",
 866 |        "      <td>701</td>\n",
 867 |        "      <td>0</td>\n",
 868 |        "      <td>2021-06-07 12:59:14.813419</td>\n",
 869 |        "    </tr>\n",
 870 |        "    <tr>\n",
 871 |        "      <th>4</th>\n",
 872 |        "      <td>xqjsd3</td>\n",
 873 |        "      <td>351</td>\n",
 874 |        "      <td>428</td>\n",
 875 |        "      <td>0</td>\n",
 876 |        "      <td>2021-06-07 12:59:14.813420</td>\n",
 877 |        "    </tr>\n",
 878 |        "    <tr>\n",
 879 |        "      <th>...</th>\n",
 880 |        "      <td>...</td>\n",
 881 |        "      <td>...</td>\n",
 882 |        "      <td>...</td>\n",
 883 |        "      <td>...</td>\n",
 884 |        "      <td>...</td>\n",
 885 |        "    </tr>\n",
 886 |        "    <tr>\n",
 887 |        "      <th>97279</th>\n",
 888 |        "      <td>h1p7lk</td>\n",
 889 |        "      <td>518</td>\n",
 890 |        "      <td>407</td>\n",
 891 |        "      <td>1</td>\n",
 892 |        "      <td>2021-06-07 12:59:14.818469</td>\n",
 893 |        "    </tr>\n",
 894 |        "    <tr>\n",
 895 |        "      <th>97325</th>\n",
 896 |        "      <td>n120dt</td>\n",
 897 |        "      <td>595</td>\n",
 898 |        "      <td>927</td>\n",
 899 |        "      <td>1</td>\n",
 900 |        "      <td>2021-06-07 12:59:14.818470</td>\n",
 901 |        "    </tr>\n",
 902 |        "    <tr>\n",
 903 |        "      <th>97818</th>\n",
 904 |        "      <td>txk4ui</td>\n",
 905 |        "      <td>583</td>\n",
 906 |        "      <td>872</td>\n",
 907 |        "      <td>1</td>\n",
 908 |        "      <td>2021-06-07 12:59:14.818470</td>\n",
 909 |        "    </tr>\n",
 910 |        "    <tr>\n",
 911 |        "      <th>98870</th>\n",
 912 |        "      <td>j72zdi</td>\n",
 913 |        "      <td>685</td>\n",
 914 |        "      <td>114</td>\n",
 915 |        "      <td>0</td>\n",
 916 |        "      <td>2021-06-07 12:59:14.818471</td>\n",
 917 |        "    </tr>\n",
 918 |        "    <tr>\n",
 919 |        "      <th>99563</th>\n",
 920 |        "      <td>wi10zj</td>\n",
 921 |        "      <td>404</td>\n",
 922 |        "      <td>627</td>\n",
 923 |        "      <td>1</td>\n",
 924 |        "      <td>2021-06-07 12:59:14.818471</td>\n",
 925 |        "    </tr>\n",
 926 |        "  </tbody>\n",
 927 |        "</table>\n",
 928 |        "<p>9944 rows × 5 columns</p>\n",
 929 |        "</div>"
 930 |       ],
 931 |       "text/plain": [
 932 |        "      user_id  credit_score  account_age_days  user_has_2fa_installed  \\\n",
 933 |        "0      782k16           626               799                       1   \n",
 934 |        "1      l0e31n           648               889                       1   \n",
 935 |        "2      mfj3xt           603               383                       1   \n",
 936 |        "3      1tyh8p           808               701                       0   \n",
 937 |        "4      xqjsd3           351               428                       0   \n",
 938 |        "...       ...           ...               ...                     ...   \n",
 939 |        "97279  h1p7lk           518               407                       1   \n",
 940 |        "97325  n120dt           595               927                       1   \n",
 941 |        "97818  txk4ui           583               872                       1   \n",
 942 |        "98870  j72zdi           685               114                       0   \n",
 943 |        "99563  wi10zj           404               627                       1   \n",
 944 |        "\n",
 945 |        "               feature_timestamp  \n",
 946 |        "0     2021-06-07 12:59:14.813413  \n",
 947 |        "1     2021-06-07 12:59:14.813418  \n",
 948 |        "2     2021-06-07 12:59:14.813419  \n",
 949 |        "3     2021-06-07 12:59:14.813419  \n",
 950 |        "4     2021-06-07 12:59:14.813420  \n",
 951 |        "...                          ...  \n",
 952 |        "97279 2021-06-07 12:59:14.818469  \n",
 953 |        "97325 2021-06-07 12:59:14.818470  \n",
 954 |        "97818 2021-06-07 12:59:14.818470  \n",
 955 |        "98870 2021-06-07 12:59:14.818471  \n",
 956 |        "99563 2021-06-07 12:59:14.818471  \n",
 957 |        "\n",
 958 |        "[9944 rows x 5 columns]"
 959 |       ]
 960 |      },
 961 |      "execution_count": 25,
 962 |      "metadata": {},
 963 |      "output_type": "execute_result"
 964 |     }
 965 |    ],
 966 |    "source": [
 967 |     "import numpy as np\n",
 968 |     "\n",
 969 |     "user_ids = t['src_account'].drop_duplicates()\n",
 970 |     "\n",
 971 |     "user_features = pd.DataFrame.from_dict({\n",
 972 |     "    \"user_id\": user_ids,\n",
 973 |     "    \"credit_score\": [int(np.random.uniform(350, 850)) % 850 for _ in range(9944)],\n",
 974 |     "    \"account_age_days\": [int(np.random.uniform(25, 960)) for _ in range(9944)],\n",
 975 |     "    \"user_has_2fa_installed\": [int(np.random.uniform(0, 2)) for _ in range(9944)],\n",
 976 |     "    \"feature_timestamp\": [datetime.now() for _ in range(9944)]\n",
 977 |     "})\n",
 978 |     "\n",
 979 |     "user_features"
 980 |    ]
 981 |   },
 982 |   {
 983 |    "cell_type": "code",
 984 |    "execution_count": 26,
 985 |    "metadata": {},
 986 |    "outputs": [
 987 |     {
 988 |      "name": "stderr",
 989 |      "output_type": "stream",
 990 |      "text": [
 991 |       "1it [00:05,  5.71s/it]\n"
 992 |      ]
 993 |     }
 994 |    ],
 995 |    "source": [
 996 |     "user_features.to_gbq(destination_table=\"fraud_tutorial.user_account_features\", project_id=\"feast-oss\", if_exists='replace')"
 997 |    ]
 998 |   },
 999 |   {
1000 |    "cell_type": "code",
1001 |    "execution_count": 125,
1002 |    "metadata": {},
1003 |    "outputs": [
1004 |     {
1005 |      "name": "stdout",
1006 |      "output_type": "stream",
1007 |      "text": [
1008 |       "1.7273118602433915\n"
1009 |      ]
1010 |     },
1011 |     {
1012 |      "data": {
1013 |       "text/plain": [
1014 |        "1"
1015 |       ]
1016 |      },
1017 |      "execution_count": 125,
1018 |      "metadata": {},
1019 |      "output_type": "execute_result"
1020 |     }
1021 |    ],
1022 |    "source": [
1023 |     "x = np.rando2021-06-04 12:56:23.851782 UTC\n",
1024 |     "m.uniform(0, 2)\n",
1025 |     "print(x)\n",
1026 |     "int(x)"
1027 |    ]
1028 |   },
1029 |   {
1030 |    "cell_type": "code",
1031 |    "execution_count": 2,
1032 |    "metadata": {},
1033 |    "outputs": [
1034 |     {
1035 |      "output_type": "stream",
1036 |      "name": "stderr",
1037 |      "text": [
1038 |       "Downloading: 100%|██████████| 100000/100000 [00:04<00:00, 20677.91rows/s]\n"
1039 |      ]
1040 |     }
1041 |    ],
1042 |    "source": [
1043 |     "import pandas as pd\n",
1044 |     "\n",
1045 |     "sql = \"\"\"\n",
1046 |     "    SELECT *\n",
1047 |     "    FROM `feast-oss.fraud_tutorial.transactions`\n",
1048 |     "\"\"\"\n",
1049 |     "\n",
1050 |     "# Run a Standard SQL query using the environment's default project\n",
1051 |     "transactions = pd.read_gbq(sql, dialect='standard')\n",
1052 |     "\n",
1053 |     "latest_time = transactions['timestamp'].max()"
1054 |    ]
1055 |   },
1056 |   {
1057 |    "cell_type": "code",
1058 |    "execution_count": 7,
1059 |    "metadata": {},
1060 |    "outputs": [
1061 |     {
1062 |      "output_type": "execute_result",
1063 |      "data": {
1064 |       "text/plain": [
1065 |        "      src_account   amount dest_account  is_fraud  \\\n",
1066 |        "0          782k16  9749.99       24bhqv         0   \n",
1067 |        "1          l0e31n  9771.14       savnzy         0   \n",
1068 |        "2          mfj3xt  4651.57       xvjv67         0   \n",
1069 |        "3          1tyh8p  8527.89       h5jgwy         0   \n",
1070 |        "4          xqjsd3  6856.42       2z3w39         0   \n",
1071 |        "...           ...      ...          ...       ...   \n",
1072 |        "99995      u49qmt  6743.81       397mqf         0   \n",
1073 |        "99996      jiaxoq  2573.43       ct01il         0   \n",
1074 |        "99997      bgf8nl  6016.44       q0ltxc         0   \n",
1075 |        "99998      yzziue  7851.83       tgx086         0   \n",
1076 |        "99999      7859ge  7078.83       waqmx5         0   \n",
1077 |        "\n",
1078 |        "                             timestamp  \n",
1079 |        "0     2021-06-10 11:53:31.513514+00:00  \n",
1080 |        "1     2021-06-10 11:53:26.483977+00:00  \n",
1081 |        "2     2021-06-10 11:53:03.224555+00:00  \n",
1082 |        "3     2021-06-10 11:53:01.317719+00:00  \n",
1083 |        "4     2021-06-10 11:52:47.205513+00:00  \n",
1084 |        "...                                ...  \n",
1085 |        "99995 2021-05-27 14:42:40.242292+00:00  \n",
1086 |        "99996 2021-05-27 14:42:17.548473+00:00  \n",
1087 |        "99997 2021-05-27 14:41:59.580318+00:00  \n",
1088 |        "99998 2021-05-27 14:41:47.836046+00:00  \n",
1089 |        "99999 2021-05-27 14:41:35.411161+00:00  \n",
1090 |        "\n",
1091 |        "[100000 rows x 5 columns]"
1092 |       ],
1093 |       "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>src_account</th>\n      <th>amount</th>\n      <th>dest_account</th>\n      <th>is_fraud</th>\n      <th>timestamp</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>782k16</td>\n      <td>9749.99</td>\n      <td>24bhqv</td>\n      <td>0</td>\n      <td>2021-06-10 11:53:31.513514+00:00</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>l0e31n</td>\n      <td>9771.14</td>\n      <td>savnzy</td>\n      <td>0</td>\n      <td>2021-06-10 11:53:26.483977+00:00</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>mfj3xt</td>\n      <td>4651.57</td>\n      <td>xvjv67</td>\n      <td>0</td>\n      <td>2021-06-10 11:53:03.224555+00:00</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>1tyh8p</td>\n      <td>8527.89</td>\n      <td>h5jgwy</td>\n      <td>0</td>\n      <td>2021-06-10 11:53:01.317719+00:00</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>xqjsd3</td>\n      <td>6856.42</td>\n      <td>2z3w39</td>\n      <td>0</td>\n      <td>2021-06-10 11:52:47.205513+00:00</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>99995</th>\n      <td>u49qmt</td>\n      <td>6743.81</td>\n      <td>397mqf</td>\n      <td>0</td>\n      <td>2021-05-27 14:42:40.242292+00:00</td>\n    </tr>\n    <tr>\n      <th>99996</th>\n      <td>jiaxoq</td>\n      <td>2573.43</td>\n      <td>ct01il</td>\n      <td>0</td>\n      <td>2021-05-27 14:42:17.548473+00:00</td>\n    </tr>\n    <tr>\n      <th>99997</th>\n      <td>bgf8nl</td>\n      <td>6016.44</td>\n      <td>q0ltxc</td>\n      <td>0</td>\n      <td>2021-05-27 14:41:59.580318+00:00</td>\n    </tr>\n    <tr>\n      <th>99998</th>\n      <td>yzziue</td>\n      <td>7851.83</td>\n      <td>tgx086</td>\n      <td>0</td>\n      <td>2021-05-27 14:41:47.836046+00:00</td>\n    </tr>\n    <tr>\n      <th>99999</th>\n      <td>7859ge</td>\n      <td>7078.83</td>\n      <td>waqmx5</td>\n      <td>0</td>\n      <td>2021-05-27 14:41:35.411161+00:00</td>\n    </tr>\n  </tbody>\n</table>\n<p>100000 rows × 5 columns</p>\n</div>"
1094 |      },
1095 |      "metadata": {},
1096 |      "execution_count": 7
1097 |     }
1098 |    ],
1099 |    "source": [
1100 |     "from datetime import datetime\n",
1101 |     "datediff = datetime.now() - latest_time.replace(tzinfo=None)\n",
1102 |     "\n",
1103 |     "transactions['timestamp'] = transactions['timestamp'] + datediff\n",
1104 |     "\n",
1105 |     "transactions"
1106 |    ]
1107 |   },
1108 |   {
1109 |    "cell_type": "code",
1110 |    "execution_count": 9,
1111 |    "metadata": {},
1112 |    "outputs": [
1113 |     {
1114 |      "output_type": "stream",
1115 |      "name": "stderr",
1116 |      "text": [
1117 |       "1it [00:09,  9.47s/it]\n"
1118 |      ]
1119 |     }
1120 |    ],
1121 |    "source": [
1122 |     "transactions.to_gbq(destination_table=\"fraud_tutorial.transactions\", project_id=\"feast-oss\", if_exists='replace')"
1123 |    ]
1124 |   },
1125 |   {
1126 |    "cell_type": "code",
1127 |    "execution_count": 5,
1128 |    "metadata": {},
1129 |    "outputs": [
1130 |     {
1131 |      "output_type": "stream",
1132 |      "name": "stderr",
1133 |      "text": [
1134 |       "Downloading: 100%|██████████| 9944/9944 [00:00<00:00, 12317.33rows/s]\n"
1135 |      ]
1136 |     },
1137 |     {
1138 |      "output_type": "execute_result",
1139 |      "data": {
1140 |       "text/plain": [
1141 |        "     user_id  credit_score  account_age_days  user_has_2fa_installed  \\\n",
1142 |        "0     41sozr           512               700                       0   \n",
1143 |        "1     h8nr8u           512               157                       0   \n",
1144 |        "2     shid6v           512               509                       0   \n",
1145 |        "3     rbcoqw           512               742                       0   \n",
1146 |        "4     hew545           512               327                       0   \n",
1147 |        "...      ...           ...               ...                     ...   \n",
1148 |        "9939  nsgtkp           767               891                       1   \n",
1149 |        "9940  4dlidj           767               855                       1   \n",
1150 |        "9941  1z87hk           767               271                       1   \n",
1151 |        "9942  ffqerm           767               829                       1   \n",
1152 |        "9943  elz674           767               783                       1   \n",
1153 |        "\n",
1154 |        "              feature_timestamp  \n",
1155 |        "0    2021-06-03 12:11:13.032174  \n",
1156 |        "1    2021-06-03 12:11:13.032174  \n",
1157 |        "2    2021-06-03 12:11:13.032174  \n",
1158 |        "3    2021-06-03 12:11:13.032174  \n",
1159 |        "4    2021-06-03 12:11:13.032174  \n",
1160 |        "...                         ...  \n",
1161 |        "9939 2021-06-03 12:11:13.032174  \n",
1162 |        "9940 2021-06-03 12:11:13.032174  \n",
1163 |        "9941 2021-06-03 12:11:13.032174  \n",
1164 |        "9942 2021-06-03 12:11:13.032174  \n",
1165 |        "9943 2021-06-03 12:11:13.032174  \n",
1166 |        "\n",
1167 |        "[9944 rows x 5 columns]"
1168 |       ],
1169 |       "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>user_id</th>\n      <th>credit_score</th>\n      <th>account_age_days</th>\n      <th>user_has_2fa_installed</th>\n      <th>feature_timestamp</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>41sozr</td>\n      <td>512</td>\n      <td>700</td>\n      <td>0</td>\n      <td>2021-06-03 12:11:13.032174</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>h8nr8u</td>\n      <td>512</td>\n      <td>157</td>\n      <td>0</td>\n      <td>2021-06-03 12:11:13.032174</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>shid6v</td>\n      <td>512</td>\n      <td>509</td>\n      <td>0</td>\n      <td>2021-06-03 12:11:13.032174</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>rbcoqw</td>\n      <td>512</td>\n      <td>742</td>\n      <td>0</td>\n      <td>2021-06-03 12:11:13.032174</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>hew545</td>\n      <td>512</td>\n      <td>327</td>\n      <td>0</td>\n      <td>2021-06-03 12:11:13.032174</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>9939</th>\n      <td>nsgtkp</td>\n      <td>767</td>\n      <td>891</td>\n      <td>1</td>\n      <td>2021-06-03 12:11:13.032174</td>\n    </tr>\n    <tr>\n      <th>9940</th>\n      <td>4dlidj</td>\n      <td>767</td>\n      <td>855</td>\n      <td>1</td>\n      <td>2021-06-03 12:11:13.032174</td>\n    </tr>\n    <tr>\n      <th>9941</th>\n      <td>1z87hk</td>\n      <td>767</td>\n      <td>271</td>\n      <td>1</td>\n      <td>2021-06-03 12:11:13.032174</td>\n    </tr>\n    <tr>\n      <th>9942</th>\n      <td>ffqerm</td>\n      <td>767</td>\n      <td>829</td>\n      <td>1</td>\n      <td>2021-06-03 12:11:13.032174</td>\n    </tr>\n    <tr>\n      <th>9943</th>\n      <td>elz674</td>\n      <td>767</td>\n      <td>783</td>\n      <td>1</td>\n      <td>2021-06-03 12:11:13.032174</td>\n    </tr>\n  </tbody>\n</table>\n<p>9944 rows × 5 columns</p>\n</div>"
1170 |      },
1171 |      "metadata": {},
1172 |      "execution_count": 5
1173 |     }
1174 |    ],
1175 |    "source": [
1176 |     "import pandas as pd\n",
1177 |     "from datetime import datetime, timedelta\n",
1178 |     "\n",
1179 |     "sql = \"\"\"\n",
1180 |     "    SELECT *\n",
1181 |     "    FROM `feast-oss.fraud_tutorial.user_account_features`\n",
1182 |     "\"\"\"\n",
1183 |     "\n",
1184 |     "user_features = pd.read_gbq(sql, dialect='standard')\n",
1185 |     "\n",
1186 |     "user_features['feature_timestamp'] = datetime.now() - timedelta(days=7)\n",
1187 |     "\n",
1188 |     "user_features"
1189 |    ]
1190 |   },
1191 |   {
1192 |    "cell_type": "code",
1193 |    "execution_count": 6,
1194 |    "metadata": {},
1195 |    "outputs": [
1196 |     {
1197 |      "output_type": "stream",
1198 |      "name": "stderr",
1199 |      "text": [
1200 |       "1it [00:04,  4.62s/it]\n"
1201 |      ]
1202 |     }
1203 |    ],
1204 |    "source": [
1205 |     "user_features.to_gbq(destination_table=\"fraud_tutorial.user_account_features\", project_id=\"feast-oss\", if_exists='replace')"
1206 |    ]
1207 |   },
1208 |   {
1209 |    "cell_type": "code",
1210 |    "execution_count": null,
1211 |    "metadata": {},
1212 |    "outputs": [],
1213 |    "source": []
1214 |   }
1215 |  ],
1216 |  "metadata": {
1217 |   "kernelspec": {
1218 |    "name": "python379jvsc74a57bd0aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49",
1219 |    "display_name": "Python 3.7.9 64-bit"
1220 |   },
1221 |   "language_info": {
1222 |    "codemirror_mode": {
1223 |     "name": "ipython",
1224 |     "version": 3
1225 |    },
1226 |    "file_extension": ".py",
1227 |    "mimetype": "text/x-python",
1228 |    "name": "python",
1229 |    "nbconvert_exporter": "python",
1230 |    "pygments_lexer": "ipython3",
1231 |    "version": "3.7.9"
1232 |   },
1233 |   "metadata": {
1234 |    "interpreter": {
1235 |     "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
1236 |    }
1237 |   }
1238 |  },
1239 |  "nbformat": 4,
1240 |  "nbformat_minor": 2
1241 | }


--------------------------------------------------------------------------------
/notebooks/Validating_Online_Features_While_Detecting_Fraud.ipynb:
--------------------------------------------------------------------------------
   1 | {
   2 |  "cells": [
   3 |   {
   4 |    "cell_type": "markdown",
   5 |    "metadata": {
   6 |     "id": "jirdTjhETQW0"
   7 |    },
   8 |    "source": [
   9 |     "# Introduction\n",
  10 |     "\n",
  11 |     "In this tutorial we will extend previously developed fraud prediction system by adding Data Quality Monitoring for online features.\n",
  12 |     "\n",
  13 |     "If you haven't already please check out these two tutorials first:\n",
  14 |     "1. [Fraud Detection (with BigQuery and Datastore)](https://github.com/feast-dev/feast-gcp-fraud-tutorial/blob/main/notebooks/Fraud_Detection_Tutorial.ipynb)\n",
  15 |     "2. [Validation of historical features with Great Expectations](https://docs.feast.dev/tutorials/validating-historical-features)\n",
  16 |     "\n",
  17 |     "Throughout this tutorial, we’ll briefly revisit set up of feature store for the fraud detection system and then we'll walk through the creation of validation expectations, configuration of the online features logging and will check how to apply validation in production.\n",
  18 |     "\n",
  19 |     "*The need to revisit the system desribed in [previous tutorial](https://github.com/feast-dev/feast-gcp-fraud-tutorial/blob/main/notebooks/Fraud_Detection_Tutorial.ipynb) is caused by the fact that Go feature server, which can produce feature logs used in validation, currently supports only Redis online store, whereas previous tutorial was using Datastore.*\n",
  20 |     "\n",
  21 |     "Here's a high-level diagram desribing data flow in DQM pipeline:\n",
  22 |     "\n",
  23 |     "\n"
  24 |    ]
  25 |   },
  26 |   {
  27 |    "cell_type": "markdown",
  28 |    "metadata": {
  29 |     "id": "2qipWwfSnrjK"
  30 |    },
  31 |    "source": [
  32 |     "![Feast landing page, blog post - Tutorial - Materialize.png](data:image/png;base64,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)"
  33 |    ]
  34 |   },
  35 |   {
  36 |    "cell_type": "markdown",
  37 |    "metadata": {},
  38 |    "source": [
  39 |     "It all starts from generating a training dataset (by pulling historical features from an offline store). This training dataset serves as a validation reference. When a data scientist develops a model some set of implicit expectations from the training dataset appears. Those expectations, first of all, should be met by the training dataset itself and only then we can apply them to the online features in production.\n",
  40 |     "\n",
  41 |     "Hence, on the second step the data scientist explores the dataset and develops or formalizes those expectations with help of [Great Expectations library](https://docs.greatexpectations.io/docs/). Those expectations can be checked right aways against training dataset and only those that pass on it will be added to a reference profile. Reference profile is a set of expectations that could be serialzed and later checked against tested dataset w/o the need to load the training dataset again.\n",
  42 |     "\n",
  43 |     "On the evaluation stage a tested dataset is loaded from a storage and validated against a reference profile."
  44 |    ]
  45 |   },
  46 |   {
  47 |    "cell_type": "markdown",
  48 |    "metadata": {},
  49 |    "source": [
  50 |     "What you'll need for this tutorial:\n",
  51 |     "1. GCP Account with access to BigQuery\n",
  52 |     "2. Redis server (accessible locally)\n",
  53 |     "3. (for Windows / Mac M1 users) installed Go compiler (=> 1.17) to build parts of Feast written in Go "
  54 |    ]
  55 |   },
  56 |   {
  57 |    "cell_type": "markdown",
  58 |    "metadata": {},
  59 |    "source": [
  60 |     "# Part I (feature store for fraud detection system on Redis)"
  61 |    ]
  62 |   },
  63 |   {
  64 |    "cell_type": "markdown",
  65 |    "metadata": {},
  66 |    "source": [
  67 |     "So let's first recall the basics of the feature store creation."
  68 |    ]
  69 |   },
  70 |   {
  71 |    "cell_type": "markdown",
  72 |    "metadata": {
  73 |     "id": "WX6daujBVgZ5"
  74 |    },
  75 |    "source": [
  76 |     "## Installation and set up\n",
  77 |     "\n",
  78 |     "### Install Feast\n",
  79 |     "\n",
  80 |     "Feast can be installed using pip. This installation includes a Python package as well as a CLI.\n",
  81 |     "\n",
  82 |     "Feast contains some packages which conflict with the default versions installed in Colab. **After running this cell, restart the runtime to continue** (Runtime > Restart runtime.)\n"
  83 |    ]
  84 |   },
  85 |   {
  86 |    "cell_type": "code",
  87 |    "execution_count": null,
  88 |    "metadata": {
  89 |     "colab": {
  90 |      "base_uri": "https://localhost:8080/"
  91 |     },
  92 |     "id": "S51NR-oPVsjg",
  93 |     "outputId": "81fa5d76-1641-4c2f-a7f7-a27988b686f8"
  94 |    },
  95 |    "outputs": [],
  96 |    "source": [
  97 |     "%env COMPILE_GO=True\n",
  98 |     "%env FEAST_USAGE=False\n",
  99 |     "\n",
 100 |     "! pip install 'feast[gcp,redis,ge,go]'\n",
 101 |     "! feast version"
 102 |    ]
 103 |   },
 104 |   {
 105 |    "cell_type": "markdown",
 106 |    "metadata": {
 107 |     "id": "6eKs1547MiFA"
 108 |    },
 109 |    "source": [
 110 |     "## Set configurations\n",
 111 |     "\n",
 112 |     "Set the following configuration, which we'll be using throughout the tutorial:\n",
 113 |     "\n",
 114 |     "- PROJECT_ID: Your project.\n",
 115 |     "- BIGQUERY_DATASET_NAME: The name of a dataset which will be used to create tables containing features and store the logs of the feature server."
 116 |    ]
 117 |   },
 118 |   {
 119 |    "cell_type": "code",
 120 |    "execution_count": null,
 121 |    "metadata": {
 122 |     "colab": {
 123 |      "base_uri": "https://localhost:8080/"
 124 |     },
 125 |     "id": "NKPT2GJ_Jb2h",
 126 |     "outputId": "06a9514b-b0fc-4aff-dd1c-6a9a62c4adae"
 127 |    },
 128 |    "outputs": [],
 129 |    "source": [
 130 |     "PROJECT_ID = \"\"\n",
 131 |     "BIGQUERY_DATASET_NAME = \"\""
 132 |    ]
 133 |   },
 134 |   {
 135 |    "cell_type": "markdown",
 136 |    "metadata": {
 137 |     "id": "hrjZkcFmZlSM"
 138 |    },
 139 |    "source": [
 140 |     "## Create a BigQuery dataset\n",
 141 |     "**Only if your dataset doesn't already exist**: Run the following cell to create your BigQuery dataset."
 142 |    ]
 143 |   },
 144 |   {
 145 |    "cell_type": "code",
 146 |    "execution_count": null,
 147 |    "metadata": {
 148 |     "colab": {
 149 |      "base_uri": "https://localhost:8080/"
 150 |     },
 151 |     "id": "_73sXuvjZzoz",
 152 |     "outputId": "31f70b1e-8eae-4099-efc1-b067eaaadf07"
 153 |    },
 154 |    "outputs": [],
 155 |    "source": [
 156 |     "! bq mk $BIGQUERY_DATASET_NAME"
 157 |    ]
 158 |   },
 159 |   {
 160 |    "cell_type": "markdown",
 161 |    "metadata": {
 162 |     "id": "3W_OsJMWkipk"
 163 |    },
 164 |    "source": [
 165 |     "## Initialize the feature repository"
 166 |    ]
 167 |   },
 168 |   {
 169 |    "cell_type": "markdown",
 170 |    "metadata": {
 171 |     "id": "Sk0fdKESD3j-"
 172 |    },
 173 |    "source": [
 174 |     "In Feast, you define your features using configuration stored in a repository. To start, initialize a feature repository."
 175 |    ]
 176 |   },
 177 |   {
 178 |    "cell_type": "code",
 179 |    "execution_count": null,
 180 |    "metadata": {
 181 |     "colab": {
 182 |      "base_uri": "https://localhost:8080/"
 183 |     },
 184 |     "id": "ASAv4kB3kkz_",
 185 |     "outputId": "d64888e0-f9ce-4b5a-a00d-a6b5c300b412"
 186 |    },
 187 |    "outputs": [],
 188 |    "source": [
 189 |     "! feast init fraud_tutorial\n",
 190 |     "%cd fraud_tutorial/\n",
 191 |     "! ls"
 192 |    ]
 193 |   },
 194 |   {
 195 |    "cell_type": "markdown",
 196 |    "metadata": {
 197 |     "id": "XDfgOvshswm5"
 198 |    },
 199 |    "source": [
 200 |     "Next, we'll edit the `feature_store.yaml` file to specify offline and online stores. Note that the `project` field in this file refers to the Feast concept of a project, not a GCP project."
 201 |    ]
 202 |   },
 203 |   {
 204 |    "cell_type": "code",
 205 |    "execution_count": null,
 206 |    "metadata": {
 207 |     "colab": {
 208 |      "base_uri": "https://localhost:8080/"
 209 |     },
 210 |     "id": "mK727g_Ik6zs",
 211 |     "outputId": "2bb6628d-4dd8-4a3c-89a3-a64da05c43d8"
 212 |    },
 213 |    "outputs": [],
 214 |    "source": [
 215 |     "feature_store = \\\n",
 216 |     "f\"\"\"project: fraud_tutorial\n",
 217 |     "registry: data/registry.db\n",
 218 |     "provider: local\n",
 219 |     "offline_store:\n",
 220 |     "  type: bigquery\n",
 221 |     "  dataset: {BIGQUERY_DATASET_NAME}\n",
 222 |     "online_store:\n",
 223 |     "  type: redis\n",
 224 |     "  connection_string: \"localhost:6379\"\n",
 225 |     "go_feature_retrieval: True\n",
 226 |     "\"\"\"\n",
 227 |     "\n",
 228 |     "with open('feature_store.yaml', \"w\") as feature_store_file:\n",
 229 |     "    feature_store_file.write(feature_store)\n",
 230 |     "\n",
 231 |     "# Print our feature_store.yaml\n",
 232 |     "! cat feature_store.yaml"
 233 |    ]
 234 |   },
 235 |   {
 236 |    "cell_type": "markdown",
 237 |    "metadata": {
 238 |     "id": "Q0dMrw4rESL7"
 239 |    },
 240 |    "source": [
 241 |     "Then, we can apply our feature repository:"
 242 |    ]
 243 |   },
 244 |   {
 245 |    "cell_type": "code",
 246 |    "execution_count": null,
 247 |    "metadata": {
 248 |     "colab": {
 249 |      "base_uri": "https://localhost:8080/"
 250 |     },
 251 |     "id": "f6QD4-lVrbdt",
 252 |     "outputId": "460f5f04-ea1c-4580-a1fe-0f755b71243e"
 253 |    },
 254 |    "outputs": [],
 255 |    "source": [
 256 |     "! feast apply"
 257 |    ]
 258 |   },
 259 |   {
 260 |    "cell_type": "markdown",
 261 |    "metadata": {
 262 |     "id": "ewiAq45-Efp0"
 263 |    },
 264 |    "source": [
 265 |     "## Creating features\n",
 266 |     "\n",
 267 |     "Next, let's make a new feature and register it to the store.\n",
 268 |     "\n",
 269 |     "This involves two steps.\n",
 270 |     "\n",
 271 |     "- **Using Bigquery**, we generate new feature values using SQL. Feast is used not to generate features, which is done in Python/SQL.\n",
 272 |     "- **Using Feast**, we register our new features in Feast by creating a FeatureView:\n",
 273 |     "\n"
 274 |    ]
 275 |   },
 276 |   {
 277 |    "cell_type": "markdown",
 278 |    "metadata": {
 279 |     "id": "0V0qK1knwjg-"
 280 |    },
 281 |    "source": [
 282 |     "## Preview the raw data"
 283 |    ]
 284 |   },
 285 |   {
 286 |    "cell_type": "code",
 287 |    "execution_count": null,
 288 |    "metadata": {},
 289 |    "outputs": [],
 290 |    "source": [
 291 |     "from google.cloud import bigquery\n",
 292 |     "bq_client = bigquery.Client(project=PROJECT_ID)"
 293 |    ]
 294 |   },
 295 |   {
 296 |    "cell_type": "code",
 297 |    "execution_count": null,
 298 |    "metadata": {
 299 |     "colab": {
 300 |      "base_uri": "https://localhost:8080/",
 301 |      "height": 459
 302 |     },
 303 |     "id": "p90F3kxgv9UO",
 304 |     "outputId": "5f540f22-3d93-4999-928f-b27292e8b62c"
 305 |    },
 306 |    "outputs": [],
 307 |    "source": [
 308 |     "j = bq_client.query(\"select * from feast-oss.fraud_tutorial.transactions limit 1000\")\n",
 309 |     "j.to_dataframe()"
 310 |    ]
 311 |   },
 312 |   {
 313 |    "cell_type": "markdown",
 314 |    "metadata": {
 315 |     "id": "3i0KOCg8wn5n"
 316 |    },
 317 |    "source": [
 318 |     "## Create a feature table using SQL"
 319 |    ]
 320 |   },
 321 |   {
 322 |    "cell_type": "markdown",
 323 |    "metadata": {
 324 |     "id": "1EGCETWpFV95"
 325 |    },
 326 |    "source": [
 327 |     "Then, run the following cell to generate features. This cell contains two functions:\n",
 328 |     "\n",
 329 |     "- `generate_user_count_features` runs a SQL query that counts the amount of transactions users have made as of a given point in time.\n",
 330 |     "\n",
 331 |     "- `backfill_features` runs this query multiple times over an interval to backfill features.\n",
 332 |     "\n"
 333 |    ]
 334 |   },
 335 |   {
 336 |    "cell_type": "code",
 337 |    "execution_count": null,
 338 |    "metadata": {
 339 |     "colab": {
 340 |      "base_uri": "https://localhost:8080/"
 341 |     },
 342 |     "id": "ZZ_z84Y1xMia",
 343 |     "outputId": "b52dcd50-3372-4d31-e52f-073854d3b21c"
 344 |    },
 345 |    "outputs": [],
 346 |    "source": [
 347 |     "from datetime import datetime, timedelta\n",
 348 |     "import time\n",
 349 |     "\n",
 350 |     "def generate_user_count_features(aggregation_end_date):\n",
 351 |     "    table_id  = f\"{PROJECT_ID}.{BIGQUERY_DATASET_NAME}.user_count_transactions_7d\"\n",
 352 |     "\n",
 353 |     "    client = bigquery.Client(project=PROJECT_ID)\n",
 354 |     "    job_config = bigquery.QueryJobConfig(destination=table_id, write_disposition='WRITE_APPEND')\n",
 355 |     "\n",
 356 |     "    aggregation_start_date = datetime.now() - timedelta(days=7)\n",
 357 |     "\n",
 358 |     "    sql = f\"\"\"\n",
 359 |     "    SELECT\n",
 360 |     "        src_account AS user_id,\n",
 361 |     "        COUNT(*) AS transaction_count_7d,\n",
 362 |     "        timestamp'{aggregation_end_date.isoformat()}' AS feature_timestamp\n",
 363 |     "    FROM\n",
 364 |     "        feast-oss.fraud_tutorial.transactions\n",
 365 |     "    WHERE\n",
 366 |     "        timestamp BETWEEN TIMESTAMP('{aggregation_start_date.isoformat()}')\n",
 367 |     "        AND TIMESTAMP('{aggregation_end_date.isoformat()}')\n",
 368 |     "    GROUP BY\n",
 369 |     "        user_id\n",
 370 |     "    \"\"\"\n",
 371 |     "\n",
 372 |     "    query_job = client.query(sql, job_config=job_config)\n",
 373 |     "    query_job.result()\n",
 374 |     "    print(f\"Generated features as of {aggregation_end_date.isoformat()}\")\n",
 375 |     "\n",
 376 |     "\n",
 377 |     "def backfill_features(earliest_aggregation_end_date, interval, num_iterations):\n",
 378 |     "    aggregation_end_date = earliest_aggregation_end_date\n",
 379 |     "    for _ in range(num_iterations):\n",
 380 |     "        generate_user_count_features(aggregation_end_date=aggregation_end_date)\n",
 381 |     "        time.sleep(1)\n",
 382 |     "        aggregation_end_date += interval\n",
 383 |     "\n",
 384 |     "if __name__ == '__main__':\n",
 385 |     "    backfill_features(\n",
 386 |     "        earliest_aggregation_end_date=datetime.now() - timedelta(days=7),\n",
 387 |     "        interval=timedelta(days=1),\n",
 388 |     "        num_iterations=8\n",
 389 |     "    )\n"
 390 |    ]
 391 |   },
 392 |   {
 393 |    "cell_type": "markdown",
 394 |    "metadata": {
 395 |     "id": "JYbQTSWiGCWu"
 396 |    },
 397 |    "source": [
 398 |     "Then, we can preview our new feature:"
 399 |    ]
 400 |   },
 401 |   {
 402 |    "cell_type": "code",
 403 |    "execution_count": null,
 404 |    "metadata": {
 405 |     "colab": {
 406 |      "base_uri": "https://localhost:8080/"
 407 |     },
 408 |     "id": "c3WpOXbUxs6d",
 409 |     "outputId": "696fa460-d89f-4286-e8c7-a53272d930e6"
 410 |    },
 411 |    "outputs": [],
 412 |    "source": [
 413 |     "j = bq_client.query(f\"select * from {BIGQUERY_DATASET_NAME}.user_count_transactions_7d limit 1000\")\n",
 414 |     "j.to_dataframe()"
 415 |    ]
 416 |   },
 417 |   {
 418 |    "cell_type": "markdown",
 419 |    "metadata": {
 420 |     "id": "RXLgVUuNGPTj"
 421 |    },
 422 |    "source": [
 423 |     "## Create a new FeatureView\n",
 424 |     "\n",
 425 |     "Create new files, `fraud_features.py`, which contains our new feature definition and `fraud_services.py`, which contains feature service definition."
 426 |    ]
 427 |   },
 428 |   {
 429 |    "cell_type": "code",
 430 |    "execution_count": null,
 431 |    "metadata": {
 432 |     "id": "F6B7Wo67yDV7"
 433 |    },
 434 |    "outputs": [],
 435 |    "source": [
 436 |     "fraud_features = \\\n",
 437 |     "f\"\"\"\n",
 438 |     "from datetime import timedelta\n",
 439 |     "from feast import BigQuerySource, FeatureView, Entity, ValueType\n",
 440 |     "\n",
 441 |     "# Add an entity for users\n",
 442 |     "user_entity = Entity(\n",
 443 |     "    name=\"user_id\",\n",
 444 |     "    description=\"A user that has executed a transaction or received a transaction\",\n",
 445 |     ")\n",
 446 |     "\n",
 447 |     "# Add a FeatureView based on our new table\n",
 448 |     "driver_stats_fv = FeatureView(\n",
 449 |     "    name=\"user_transaction_count_7d\",\n",
 450 |     "    entities=[user_entity],\n",
 451 |     "    ttl=timedelta(weeks=1),\n",
 452 |     "    batch_source=BigQuerySource(\n",
 453 |     "        table=f\"{PROJECT_ID}.{BIGQUERY_DATASET_NAME}.user_count_transactions_7d\",\n",
 454 |     "        timestamp_field=\"feature_timestamp\"))\n",
 455 |     "\n",
 456 |     "# Add two FeatureViews based on existing tables in BigQuery\n",
 457 |     "user_account_fv = FeatureView(\n",
 458 |     "    name=\"user_account_features\",\n",
 459 |     "    entities=[user_entity],\n",
 460 |     "    ttl=timedelta(weeks=52),\n",
 461 |     "    batch_source=BigQuerySource(\n",
 462 |     "        table=f\"feast-oss.fraud_tutorial.user_account_features\",\n",
 463 |     "        timestamp_field=\"feature_timestamp\"))\n",
 464 |     "\n",
 465 |     "user_has_fraudulent_transactions_fv = FeatureView(\n",
 466 |     "    name=\"user_has_fraudulent_transactions\",\n",
 467 |     "    entities=[user_entity],\n",
 468 |     "    ttl=timedelta(weeks=52),\n",
 469 |     "    batch_source=BigQuerySource(\n",
 470 |     "        table=f\"feast-oss.fraud_tutorial.user_has_fraudulent_transactions\",\n",
 471 |     "        timestamp_field=\"feature_timestamp\"))\n",
 472 |     "\"\"\"\n",
 473 |     "\n",
 474 |     "fraud_services = f\"\"\"\n",
 475 |     "from feast import FeatureService\n",
 476 |     "\n",
 477 |     "from fraud_features import driver_stats_fv, user_account_fv, user_has_fraudulent_transactions_fv\n",
 478 |     "\n",
 479 |     "fs = FeatureService(\n",
 480 |     "    name=\"user_features\",\n",
 481 |     "    features=[\n",
 482 |     "        driver_stats_fv[[\"user_transaction_count_7d\"]],\n",
 483 |     "        user_account_fv[[\"credit_score\", \"account_age_days\", \"user_has_2fa_installed\"]],\n",
 484 |     "        user_has_fraudulent_transactions_fv[[\"user_has_fraudulent_transactions_7d\"]],\n",
 485 |     "    ],\n",
 486 |     ")\"\"\"\n",
 487 |     "\n",
 488 |     "with open('fraud_features.py', \"w\") as fraud_features_file:\n",
 489 |     "    fraud_features_file.write(fraud_features)\n",
 490 |     "    \n",
 491 |     "with open('fraud_services.py', \"w\") as fraud_services_file:\n",
 492 |     "    fraud_services_file.write(fraud_services)"
 493 |    ]
 494 |   },
 495 |   {
 496 |    "cell_type": "code",
 497 |    "execution_count": null,
 498 |    "metadata": {
 499 |     "colab": {
 500 |      "base_uri": "https://localhost:8080/"
 501 |     },
 502 |     "id": "6rSndNQ30ASP",
 503 |     "outputId": "84e0d430-d7a0-4b87-aa15-86d0cdccaf9e"
 504 |    },
 505 |    "outputs": [],
 506 |    "source": [
 507 |     "# Remove example features\n",
 508 |     "!rm example.py\n",
 509 |     "# Apply our changes\n",
 510 |     "!feast apply"
 511 |    ]
 512 |   },
 513 |   {
 514 |    "cell_type": "markdown",
 515 |    "metadata": {
 516 |     "id": "6cJFAJiuGxM3"
 517 |    },
 518 |    "source": [
 519 |     "## Fetching training data\n",
 520 |     "\n",
 521 |     "Now that our feature is registered in Feast, we can use Feast to generate a training dataset. To do this, we need an entity dataframe, alongside the list of features we want:"
 522 |    ]
 523 |   },
 524 |   {
 525 |    "cell_type": "code",
 526 |    "execution_count": null,
 527 |    "metadata": {
 528 |     "colab": {
 529 |      "base_uri": "https://localhost:8080/",
 530 |      "height": 330
 531 |     },
 532 |     "id": "BqgiEP2Oz42q",
 533 |     "outputId": "40552317-644b-4ee6-d735-0ee5e48e79dd"
 534 |    },
 535 |    "outputs": [],
 536 |    "source": [
 537 |     "from datetime import datetime, timedelta\n",
 538 |     "from feast import FeatureStore\n",
 539 |     "\n",
 540 |     "# Initialize a FeatureStore with our current repository's configurations\n",
 541 |     "store = FeatureStore(repo_path=\".\")\n",
 542 |     "\n",
 543 |     "# Get training data\n",
 544 |     "now = datetime.now()\n",
 545 |     "two_days_ago = datetime.now() - timedelta(days=2)\n",
 546 |     "\n",
 547 |     "feature_service = store.get_feature_service(\"user_features\")\n",
 548 |     "\n",
 549 |     "training_data_job = store.get_historical_features(\n",
 550 |     "    entity_df=f\"\"\"\n",
 551 |     "    select \n",
 552 |     "        src_account as user_id,\n",
 553 |     "        timestamp as event_timestamp,\n",
 554 |     "        is_fraud\n",
 555 |     "    from\n",
 556 |     "        feast-oss.fraud_tutorial.transactions\n",
 557 |     "    where\n",
 558 |     "        timestamp between timestamp('{two_days_ago.isoformat()}') \n",
 559 |     "        and timestamp('{now.isoformat()}')\"\"\",\n",
 560 |     "    features=feature_service,\n",
 561 |     "    full_feature_names=True\n",
 562 |     ")\n",
 563 |     "\n",
 564 |     "training_data = training_data_job.to_df()\n",
 565 |     "training_data.head()\n"
 566 |    ]
 567 |   },
 568 |   {
 569 |    "cell_type": "markdown",
 570 |    "metadata": {
 571 |     "id": "3izkr_3sG1hX"
 572 |    },
 573 |    "source": [
 574 |     "## Training a model\n",
 575 |     "\n",
 576 |     "Now, we can use our features to train a model:"
 577 |    ]
 578 |   },
 579 |   {
 580 |    "cell_type": "code",
 581 |    "execution_count": null,
 582 |    "metadata": {},
 583 |    "outputs": [],
 584 |    "source": [
 585 |     "!pip install sklearn"
 586 |    ]
 587 |   },
 588 |   {
 589 |    "cell_type": "code",
 590 |    "execution_count": null,
 591 |    "metadata": {
 592 |     "colab": {
 593 |      "base_uri": "https://localhost:8080/"
 594 |     },
 595 |     "id": "YMgFuFHR4pIu",
 596 |     "outputId": "7f248543-3016-482c-9d8e-b0819cf3609d"
 597 |    },
 598 |    "outputs": [],
 599 |    "source": [
 600 |     "from sklearn.linear_model import LinearRegression\n",
 601 |     "\n",
 602 |     "# Drop stray nulls\n",
 603 |     "training_data.dropna(inplace=True)\n",
 604 |     "\n",
 605 |     "# Select training matrices\n",
 606 |     "X = training_data[[\n",
 607 |     "    \"user_transaction_count_7d__transaction_count_7d\", \n",
 608 |     "    \"user_account_features__credit_score\",\n",
 609 |     "    \"user_account_features__account_age_days\",\n",
 610 |     "    \"user_account_features__user_has_2fa_installed\",\n",
 611 |     "    \"user_has_fraudulent_transactions__user_has_fraudulent_transactions_7d\"\n",
 612 |     "]]\n",
 613 |     "y = training_data[\"is_fraud\"]\n",
 614 |     "\n",
 615 |     "# Train a simple SVC model\n",
 616 |     "model = LinearRegression()\n",
 617 |     "model.fit(X, y)"
 618 |    ]
 619 |   },
 620 |   {
 621 |    "cell_type": "code",
 622 |    "execution_count": null,
 623 |    "metadata": {
 624 |     "colab": {
 625 |      "base_uri": "https://localhost:8080/"
 626 |     },
 627 |     "id": "PobCpVWu4pdz",
 628 |     "outputId": "60a39792-d332-4edc-c59c-e4e20a87ea5e"
 629 |    },
 630 |    "outputs": [],
 631 |    "source": [
 632 |     "# Get first two rows of training data\n",
 633 |     "samples = X.iloc[:2]\n",
 634 |     "\n",
 635 |     "# Make a test prediction\n",
 636 |     "model.predict(samples)"
 637 |    ]
 638 |   },
 639 |   {
 640 |    "cell_type": "code",
 641 |    "execution_count": null,
 642 |    "metadata": {},
 643 |    "outputs": [],
 644 |    "source": [
 645 |     "import joblib\n",
 646 |     "joblib.dump(model, \"model.bin\")"
 647 |    ]
 648 |   },
 649 |   {
 650 |    "cell_type": "markdown",
 651 |    "metadata": {
 652 |     "id": "HWQLT0gTHi1h"
 653 |    },
 654 |    "source": [
 655 |     "## Materializing features\n",
 656 |     "\n",
 657 |     "To enable real time feature inference, Feast loads your features into a key-value store so they're available at low latency. We use Redis as this key-value store."
 658 |    ]
 659 |   },
 660 |   {
 661 |    "cell_type": "code",
 662 |    "execution_count": null,
 663 |    "metadata": {
 664 |     "id": "REBqJPcZ99Lj"
 665 |    },
 666 |    "outputs": [],
 667 |    "source": [
 668 |     "!feast materialize-incremental $(date -u +\"%Y-%m-%dT%H:%M:%S\")"
 669 |    ]
 670 |   },
 671 |   {
 672 |    "cell_type": "markdown",
 673 |    "metadata": {
 674 |     "id": "FNVxJpw1HohD"
 675 |    },
 676 |    "source": [
 677 |     "## Low latency inference\n",
 678 |     "\n",
 679 |     "To make a prediction in real-time, we need to do the following:\n",
 680 |     "\n",
 681 |     "1. Start a feature server (as a subprocess) that will expose gRPC API\n",
 682 |     "2. Create a gRPC client using precompiled Feast Serving proto interface\n",
 683 |     "3. Call GetOnlineFeatures on this gRPC client\n",
 684 |     "4. Pass these features to the model\n",
 685 |     "5. Return these predictions to the user"
 686 |    ]
 687 |   },
 688 |   {
 689 |    "cell_type": "code",
 690 |    "execution_count": null,
 691 |    "metadata": {},
 692 |    "outputs": [],
 693 |    "source": [
 694 |     "import subprocess\n",
 695 |     "\n",
 696 |     "server = subprocess.Popen([\"feast\", \"serve\", \"-t\", \"grpc\"])"
 697 |    ]
 698 |   },
 699 |   {
 700 |    "cell_type": "code",
 701 |    "execution_count": null,
 702 |    "metadata": {},
 703 |    "outputs": [],
 704 |    "source": [
 705 |     "import grpc\n",
 706 |     "\n",
 707 |     "from feast.protos.feast.serving.ServingService_pb2 import GetOnlineFeaturesRequest\n",
 708 |     "from feast.protos.feast.serving.ServingService_pb2_grpc import ServingServiceStub\n",
 709 |     "\n",
 710 |     "from feast.protos.feast.types.Value_pb2 import RepeatedValue\n",
 711 |     "\n",
 712 |     "from feast.type_map import python_values_to_proto_values\n",
 713 |     "from feast.online_response import OnlineResponse\n",
 714 |     "\n",
 715 |     "chan = grpc.insecure_channel(\"localhost:6566\")\n",
 716 |     "grpc_client = ServingServiceStub(chan)\n",
 717 |     "\n",
 718 |     "def get_online_features_remote(user_ids):\n",
 719 |     "    resp = grpc_client.GetOnlineFeatures(\n",
 720 |     "        GetOnlineFeaturesRequest(\n",
 721 |     "            feature_service=\"user_features\",\n",
 722 |     "            entities={\n",
 723 |     "                \"user_id\": RepeatedValue(\n",
 724 |     "                    val=python_values_to_proto_values(user_ids)\n",
 725 |     "                )\n",
 726 |     "            }\n",
 727 |     "        )\n",
 728 |     "    )\n",
 729 |     "    return OnlineResponse(resp).to_dict()"
 730 |    ]
 731 |   },
 732 |   {
 733 |    "cell_type": "code",
 734 |    "execution_count": null,
 735 |    "metadata": {
 736 |     "colab": {
 737 |      "base_uri": "https://localhost:8080/"
 738 |     },
 739 |     "id": "JyME-zoy-4d_",
 740 |     "outputId": "b2e9ba19-4d0b-4cac-9c69-90a06c28fe25"
 741 |    },
 742 |    "outputs": [],
 743 |    "source": [
 744 |     "import joblib\n",
 745 |     "model = joblib.load('model.bin')\n",
 746 |     "\n",
 747 |     "def predict(user_ids):\n",
 748 |     "    feature_vector = get_online_features_remote(user_ids)\n",
 749 |     "    \n",
 750 |     "    # Delete entity keys\n",
 751 |     "    del feature_vector[\"user_id\"]\n",
 752 |     "\n",
 753 |     "    # Flatten response from Feast\n",
 754 |     "    instances = [\n",
 755 |     "        [feature_values[i] or 0 for feature_values in feature_vector.values()]\n",
 756 |     "        for i in range(len(user_ids))\n",
 757 |     "    ]\n",
 758 |     "\n",
 759 |     "    response = model.predict(instances)\n",
 760 |     "    return response\n",
 761 |     "\n",
 762 |     "predict([\"v5zlw0\"])"
 763 |    ]
 764 |   },
 765 |   {
 766 |    "cell_type": "markdown",
 767 |    "metadata": {},
 768 |    "source": [
 769 |     " "
 770 |    ]
 771 |   },
 772 |   {
 773 |    "cell_type": "markdown",
 774 |    "metadata": {},
 775 |    "source": [
 776 |     " "
 777 |    ]
 778 |   },
 779 |   {
 780 |    "cell_type": "markdown",
 781 |    "metadata": {},
 782 |    "source": [
 783 |     "# Part II (online features logging and validation)"
 784 |    ]
 785 |   },
 786 |   {
 787 |    "cell_type": "markdown",
 788 |    "metadata": {},
 789 |    "source": [
 790 |     "In this part we will extend our feature store project with Data Quality Monitoring. Specifically, we are going to validate online features (features served by the feature server) against the reference dataset created from training features by applying expectations, that we are going to develop ourselves. We will do this in 3 steps:\n",
 791 |     "1. Configuring feature logging in the feature server and setting destination for the specific feature service object.\n",
 792 |     "2. Defining expectations using [Great Expectations](https://greatexpectations.io/).\n",
 793 |     "3. Triggering validation using SDK or CLI API."
 794 |    ]
 795 |   },
 796 |   {
 797 |    "cell_type": "markdown",
 798 |    "metadata": {},
 799 |    "source": [
 800 |     "### Updating configuration to enable logging"
 801 |    ]
 802 |   },
 803 |   {
 804 |    "cell_type": "markdown",
 805 |    "metadata": {},
 806 |    "source": [
 807 |     "First, let's edit our `feature_store.yaml` and add `feature_logging` parameter inside `feature_server`."
 808 |    ]
 809 |   },
 810 |   {
 811 |    "cell_type": "code",
 812 |    "execution_count": null,
 813 |    "metadata": {},
 814 |    "outputs": [],
 815 |    "source": [
 816 |     "feature_store = \\\n",
 817 |     "f\"\"\"project: fraud_tutorial\n",
 818 |     "registry: data/registry.db\n",
 819 |     "provider: local\n",
 820 |     "offline_store:\n",
 821 |     "  type: bigquery\n",
 822 |     "  dataset: {BIGQUERY_DATASET_NAME}\n",
 823 |     "online_store:\n",
 824 |     "  type: redis\n",
 825 |     "  connection_string: \"localhost:6379\"\n",
 826 |     "feature_server:\n",
 827 |     "  enabled: True\n",
 828 |     "  feature_logging:\n",
 829 |     "    enabled: True\n",
 830 |     "    flush_interval_secs: 60\n",
 831 |     "    write_to_disk_interval_secs: 10\n",
 832 |     "    \n",
 833 |     "go_feature_retrieval: True\n",
 834 |     "\"\"\"\n",
 835 |     "\n",
 836 |     "with open('feature_store.yaml', \"w\") as feature_store_file:\n",
 837 |     "    feature_store_file.write(feature_store)"
 838 |    ]
 839 |   },
 840 |   {
 841 |    "cell_type": "markdown",
 842 |    "metadata": {},
 843 |    "source": [
 844 |     "Next, we need to update our feature service definition in `fraud_services.py` with `logging config`. Logging config defines sample rate and logging destination. Sample rate sets the percentage of feature rows that will be logged from all served features and can have a value from 0 to 1 (inclusive from both sides). Destination must be a table or a path in the offline store declared globally in the `feature_store.yaml`."
 845 |    ]
 846 |   },
 847 |   {
 848 |    "cell_type": "code",
 849 |    "execution_count": null,
 850 |    "metadata": {},
 851 |    "outputs": [],
 852 |    "source": [
 853 |     "fraud_services = f\"\"\"\n",
 854 |     "from feast import FeatureService\n",
 855 |     "from feast.feature_logging import LoggingConfig\n",
 856 |     "from feast.infra.offline_stores.bigquery_source import BigQueryLoggingDestination\n",
 857 |     "\n",
 858 |     "from fraud_features import driver_stats_fv, user_account_fv, user_has_fraudulent_transactions_fv\n",
 859 |     "\n",
 860 |     "fs = FeatureService(\n",
 861 |     "    name=\"user_features\",\n",
 862 |     "    features=[\n",
 863 |     "        driver_stats_fv[[\"user_transaction_count_7d\"]],\n",
 864 |     "        user_account_fv[[\"credit_score\", \"account_age_days\", \"user_has_2fa_installed\"]],\n",
 865 |     "        user_has_fraudulent_transactions_fv[[\"user_has_fraudulent_transactions_7d\"]],\n",
 866 |     "    ],\n",
 867 |     "    logging_config=LoggingConfig(\n",
 868 |     "        sample_rate=1.0,\n",
 869 |     "        destination=BigQueryLoggingDestination(\n",
 870 |     "            table_ref=\"{PROJECT_ID}.{BIGQUERY_DATASET_NAME}.user_features_online_logs\"\n",
 871 |     "        )\n",
 872 |     "    )\n",
 873 |     ")\"\"\"\n",
 874 |     "    \n",
 875 |     "with open('fraud_services.py', \"w\") as fraud_services_file:\n",
 876 |     "    fraud_services_file.write(fraud_services)"
 877 |    ]
 878 |   },
 879 |   {
 880 |    "cell_type": "markdown",
 881 |    "metadata": {},
 882 |    "source": [
 883 |     "Now, let's apply these changes:"
 884 |    ]
 885 |   },
 886 |   {
 887 |    "cell_type": "code",
 888 |    "execution_count": null,
 889 |    "metadata": {},
 890 |    "outputs": [],
 891 |    "source": [
 892 |     "! feast apply"
 893 |    ]
 894 |   },
 895 |   {
 896 |    "cell_type": "markdown",
 897 |    "metadata": {},
 898 |    "source": [
 899 |     "and restart the feature server:"
 900 |    ]
 901 |   },
 902 |   {
 903 |    "cell_type": "code",
 904 |    "execution_count": null,
 905 |    "metadata": {},
 906 |    "outputs": [],
 907 |    "source": [
 908 |     "server.terminate()\n",
 909 |     "server = subprocess.Popen([\"feast\", \"serve\", \"-t\", \"grpc\"])"
 910 |    ]
 911 |   },
 912 |   {
 913 |    "cell_type": "markdown",
 914 |    "metadata": {},
 915 |    "source": [
 916 |     "### Creating reference dataset from training features"
 917 |    ]
 918 |   },
 919 |   {
 920 |    "cell_type": "code",
 921 |    "execution_count": null,
 922 |    "metadata": {},
 923 |    "outputs": [],
 924 |    "source": [
 925 |     "from feast.infra.offline_stores.bigquery_source import SavedDatasetBigQueryStorage\n",
 926 |     "\n",
 927 |     "reference_dataset = store.create_saved_dataset(\n",
 928 |     "    from_=training_data_job,\n",
 929 |     "    name=\"reference_dataset\",\n",
 930 |     "    storage=SavedDatasetBigQueryStorage(table=f\"{PROJECT_ID}.{BIGQUERY_DATASET_NAME}.reference_dataset\"))"
 931 |    ]
 932 |   },
 933 |   {
 934 |    "cell_type": "markdown",
 935 |    "metadata": {},
 936 |    "source": [
 937 |     "### Creating & testing validation profiler"
 938 |    ]
 939 |   },
 940 |   {
 941 |    "cell_type": "code",
 942 |    "execution_count": null,
 943 |    "metadata": {},
 944 |    "outputs": [],
 945 |    "source": [
 946 |     "from feast.dqm.profilers.ge_profiler import ge_profiler\n",
 947 |     "from great_expectations.dataset import PandasDataset\n",
 948 |     "from great_expectations.core.expectation_suite import ExpectationSuite"
 949 |    ]
 950 |   },
 951 |   {
 952 |    "cell_type": "markdown",
 953 |    "metadata": {},
 954 |    "source": [
 955 |     "Profiler is defined as a function that takes a dataset, a Pandas DataFrame wrapped into GE's `PandasDataset` class, and returns `ExpectationSuite`, a set of expectations:"
 956 |    ]
 957 |   },
 958 |   {
 959 |    "cell_type": "code",
 960 |    "execution_count": null,
 961 |    "metadata": {},
 962 |    "outputs": [],
 963 |    "source": [
 964 |     "@ge_profiler\n",
 965 |     "def user_features_profiler(ds: PandasDataset) -> ExpectationSuite:\n",
 966 |     "    ds.expect_column_values_to_be_between(\"user_account_features__credit_score\", 300, 850)\n",
 967 |     "    ds.expect_column_values_to_be_between(\"user_transaction_count_7d__transaction_count_7d\", min_value=0)\n",
 968 |     "    return ds.get_expectation_suite()"
 969 |    ]
 970 |   },
 971 |   {
 972 |    "cell_type": "markdown",
 973 |    "metadata": {},
 974 |    "source": [
 975 |     "To learn more about expectation functions that can be used in the profiler definition please refer to [Great Expectations doc](https://docs.greatexpectations.io/docs/)."
 976 |    ]
 977 |   },
 978 |   {
 979 |    "cell_type": "markdown",
 980 |    "metadata": {},
 981 |    "source": [
 982 |     "Profiler can be tested using the saved dataset object created above:"
 983 |    ]
 984 |   },
 985 |   {
 986 |    "cell_type": "code",
 987 |    "execution_count": null,
 988 |    "metadata": {},
 989 |    "outputs": [],
 990 |    "source": [
 991 |     "reference_dataset.get_profile(profiler=user_features_profiler)"
 992 |    ]
 993 |   },
 994 |   {
 995 |    "cell_type": "markdown",
 996 |    "metadata": {},
 997 |    "source": [
 998 |     "Profiler function along with the reference dataset must be stored in the Feast registry before calling validation API:"
 999 |    ]
1000 |   },
1001 |   {
1002 |    "cell_type": "code",
1003 |    "execution_count": null,
1004 |    "metadata": {},
1005 |    "outputs": [],
1006 |    "source": [
1007 |     "from feast.saved_dataset import ValidationReference\n",
1008 |     "\n",
1009 |     "ref = ValidationReference(\n",
1010 |     "    name='user_features_training_ref',\n",
1011 |     "    dataset_name=\"reference_dataset\",\n",
1012 |     "    profiler=user_features_profiler,\n",
1013 |     ")\n",
1014 |     "store.apply(ref)"
1015 |    ]
1016 |   },
1017 |   {
1018 |    "cell_type": "markdown",
1019 |    "metadata": {},
1020 |    "source": [
1021 |     "## Validation"
1022 |    ]
1023 |   },
1024 |   {
1025 |    "cell_type": "markdown",
1026 |    "metadata": {},
1027 |    "source": [
1028 |     "Let's now run a few predictions to log some data points:"
1029 |    ]
1030 |   },
1031 |   {
1032 |    "cell_type": "code",
1033 |    "execution_count": null,
1034 |    "metadata": {},
1035 |    "outputs": [],
1036 |    "source": [
1037 |     "user_ids = list(training_data.user_id.sample(10))\n",
1038 |     "predict(user_ids)"
1039 |    ]
1040 |   },
1041 |   {
1042 |    "cell_type": "markdown",
1043 |    "metadata": {},
1044 |    "source": [
1045 |     "After some time passed (depending on the value of `flush_interval_secs` defined in `feature_store.yaml`) we can trigger a validation:"
1046 |    ]
1047 |   },
1048 |   {
1049 |    "cell_type": "code",
1050 |    "execution_count": null,
1051 |    "metadata": {},
1052 |    "outputs": [],
1053 |    "source": [
1054 |     "end_ts = datetime.now()\n",
1055 |     "start_ts = end_ts - timedelta(minutes=10)\n",
1056 |     "\n",
1057 |     "! feast validate --feature-service user_features \\\n",
1058 |     "    --reference user_features_training_ref {start_ts.isoformat()} {end_ts.isoformat()}"
1059 |    ]
1060 |   },
1061 |   {
1062 |    "cell_type": "markdown",
1063 |    "metadata": {},
1064 |    "source": [
1065 |     "### Making validation fail"
1066 |    ]
1067 |   },
1068 |   {
1069 |    "cell_type": "markdown",
1070 |    "metadata": {},
1071 |    "source": [
1072 |     "Now, if some invalid data, that doesn't met our expectations, will be ingested into the online store and then retrieved via the feature server we should observe how validation is failing."
1073 |    ]
1074 |   },
1075 |   {
1076 |    "cell_type": "code",
1077 |    "execution_count": null,
1078 |    "metadata": {},
1079 |    "outputs": [],
1080 |    "source": [
1081 |     "import pandas as pd\n",
1082 |     "insert_df = pd.DataFrame({\n",
1083 |     "    \"user_id\": [\"pwvabf\"],\n",
1084 |     "    \"transaction_count_7d\": [-1],\n",
1085 |     "    \"feature_timestamp\": [datetime.now()],\n",
1086 |     "})\n",
1087 |     "store.write_to_online_store(\"user_transaction_count_7d\", insert_df)"
1088 |    ]
1089 |   },
1090 |   {
1091 |    "cell_type": "code",
1092 |    "execution_count": null,
1093 |    "metadata": {},
1094 |    "outputs": [],
1095 |    "source": [
1096 |     "predict([\"pwvabf\"])"
1097 |    ]
1098 |   },
1099 |   {
1100 |    "cell_type": "markdown",
1101 |    "metadata": {},
1102 |    "source": [
1103 |     "*Remember that it takes some time to write logs to BigQuery*"
1104 |    ]
1105 |   },
1106 |   {
1107 |    "cell_type": "code",
1108 |    "execution_count": null,
1109 |    "metadata": {},
1110 |    "outputs": [],
1111 |    "source": [
1112 |     "end_ts = datetime.now()\n",
1113 |     "start_ts = end_ts - timedelta(minutes=10)\n",
1114 |     "\n",
1115 |     "! feast validate --feature-service user_features \\\n",
1116 |     "    --reference user_features_training_ref {start_ts.isoformat()} {end_ts.isoformat()}"
1117 |    ]
1118 |   },
1119 |   {
1120 |    "cell_type": "markdown",
1121 |    "metadata": {},
1122 |    "source": [
1123 |     "### Alternative example with validating feature presence"
1124 |    ]
1125 |   },
1126 |   {
1127 |    "cell_type": "markdown",
1128 |    "metadata": {},
1129 |    "source": [
1130 |     "In this example we create an expectation that the feature will have a not-null value in 99% of the cases:"
1131 |    ]
1132 |   },
1133 |   {
1134 |    "cell_type": "code",
1135 |    "execution_count": null,
1136 |    "metadata": {},
1137 |    "outputs": [],
1138 |    "source": [
1139 |     "@ge_profiler\n",
1140 |     "def user_features_profiler_v2(ds: PandasDataset) -> ExpectationSuite:\n",
1141 |     "    ds.expect_column_values_to_not_be_null(\"user_account_features__account_age_days\", mostly=0.99)\n",
1142 |     "    return ds.get_expectation_suite()"
1143 |    ]
1144 |   },
1145 |   {
1146 |    "cell_type": "markdown",
1147 |    "metadata": {},
1148 |    "source": [
1149 |     "testing on the reference dataset:"
1150 |    ]
1151 |   },
1152 |   {
1153 |    "cell_type": "code",
1154 |    "execution_count": null,
1155 |    "metadata": {},
1156 |    "outputs": [],
1157 |    "source": [
1158 |     "reference_dataset.get_profile(profiler=user_features_profiler_v2)"
1159 |    ]
1160 |   },
1161 |   {
1162 |    "cell_type": "markdown",
1163 |    "metadata": {},
1164 |    "source": [
1165 |     ".. and storing new validation reference in the registry:"
1166 |    ]
1167 |   },
1168 |   {
1169 |    "cell_type": "code",
1170 |    "execution_count": null,
1171 |    "metadata": {},
1172 |    "outputs": [],
1173 |    "source": [
1174 |     "store.apply(\n",
1175 |     "    ValidationReference(\n",
1176 |     "        name='user_features_training_ref_v2',\n",
1177 |     "        dataset_name=\"reference_dataset\",\n",
1178 |     "        profiler=user_features_profiler_v2,\n",
1179 |     "    )\n",
1180 |     ")"
1181 |    ]
1182 |   },
1183 |   {
1184 |    "cell_type": "markdown",
1185 |    "metadata": {},
1186 |    "source": [
1187 |     "Retrieving some entity rows that do not exist in the online store (and thus, returned feature statuses will be NOT FOUND):"
1188 |    ]
1189 |   },
1190 |   {
1191 |    "cell_type": "code",
1192 |    "execution_count": null,
1193 |    "metadata": {},
1194 |    "outputs": [],
1195 |    "source": [
1196 |     "predict([\"invalid\"] * 5)"
1197 |    ]
1198 |   },
1199 |   {
1200 |    "cell_type": "markdown",
1201 |    "metadata": {},
1202 |    "source": [
1203 |     "Now validation should fail:"
1204 |    ]
1205 |   },
1206 |   {
1207 |    "cell_type": "code",
1208 |    "execution_count": null,
1209 |    "metadata": {},
1210 |    "outputs": [],
1211 |    "source": [
1212 |     "end_ts = datetime.now()\n",
1213 |     "start_ts = end_ts - timedelta(hours=1)\n",
1214 |     "\n",
1215 |     "! feast validate --feature-service user_features \\\n",
1216 |     "    --reference user_features_training_ref_v2 {start_ts.isoformat()} {end_ts.isoformat()}"
1217 |    ]
1218 |   },
1219 |   {
1220 |    "cell_type": "markdown",
1221 |    "metadata": {
1222 |     "id": "m4Pu2m4KUrbp"
1223 |    },
1224 |    "source": [
1225 |     "# Cleanup\n",
1226 |     "\n",
1227 |     "If you want to clean up the resources created during this tutorial, run the following cells:\n"
1228 |    ]
1229 |   },
1230 |   {
1231 |    "cell_type": "code",
1232 |    "execution_count": null,
1233 |    "metadata": {
1234 |     "colab": {
1235 |      "base_uri": "https://localhost:8080/"
1236 |     },
1237 |     "id": "9RK_Kxj2VFQu",
1238 |     "outputId": "c6947143-d41b-4ce8-9f14-36c530234eb4"
1239 |    },
1240 |    "outputs": [],
1241 |    "source": [
1242 |     "!bq rm -t -f ${BIGQUERY_DATASET_NAME}.user_count_transactions_7d\n",
1243 |     "!bq rm -t -f ${BIGQUERY_DATASET_NAME}.user_features_online_logs\n",
1244 |     "!bq rm -r -f -d ${BIGQUERY_DATASET_NAME}"
1245 |    ]
1246 |   },
1247 |   {
1248 |    "cell_type": "code",
1249 |    "execution_count": null,
1250 |    "metadata": {
1251 |     "id": "5EGuKSupu5jN"
1252 |    },
1253 |    "outputs": [],
1254 |    "source": [
1255 |     "server.terminate()"
1256 |    ]
1257 |   },
1258 |   {
1259 |    "cell_type": "code",
1260 |    "execution_count": null,
1261 |    "metadata": {},
1262 |    "outputs": [],
1263 |    "source": []
1264 |   }
1265 |  ],
1266 |  "metadata": {
1267 |   "colab": {
1268 |    "collapsed_sections": [],
1269 |    "name": "Fraud_Detection_Tutorial.ipynb",
1270 |    "provenance": [],
1271 |    "toc_visible": true
1272 |   },
1273 |   "kernelspec": {
1274 |    "display_name": "Python 3 (ipykernel)",
1275 |    "language": "python",
1276 |    "name": "python3"
1277 |   },
1278 |   "language_info": {
1279 |    "codemirror_mode": {
1280 |     "name": "ipython",
1281 |     "version": 3
1282 |    },
1283 |    "file_extension": ".py",
1284 |    "mimetype": "text/x-python",
1285 |    "name": "python",
1286 |    "nbconvert_exporter": "python",
1287 |    "pygments_lexer": "ipython3",
1288 |    "version": "3.9.12"
1289 |   }
1290 |  },
1291 |  "nbformat": 4,
1292 |  "nbformat_minor": 1
1293 | }
1294 | 


--------------------------------------------------------------------------------