├── .gitignore ├── LICENSE ├── README.rst ├── demo ├── mkdata.py ├── reactor │ ├── allproc.sls │ ├── event.yml │ └── master └── salt │ └── beacons │ └── allproc.py ├── docs ├── Makefile ├── conf.py ├── index.rst ├── make.bat └── topics │ ├── data.rst │ ├── egress.rst │ ├── flows.rst │ ├── ingress.rst │ ├── models.rst │ ├── quickstart.rst │ └── salt.rst ├── flow.yml ├── requirements.txt ├── setup.py └── umbra ├── conf.py ├── data ├── init.py └── salt_event.py ├── egress ├── cli.py ├── init.py └── salt_event.py ├── flows └── init.py ├── ingress ├── init.py ├── json_file.py └── salt_event.py ├── models ├── auto_encoder.py ├── hbos.py ├── init.py ├── kmeans.py ├── knn.py ├── lof.py ├── lscp.py └── optics.py ├── persist ├── init.py └── msgpack_file.py ├── scripts.py ├── umbra └── init.py └── version.py /.gitignore: -------------------------------------------------------------------------------- 1 | # Byte-compiled / optimized / DLL files 2 | __pycache__/ 3 | *.py[cod] 4 | *$py.class 5 | 6 | # C extensions 7 | *.so 8 | 9 | # Distribution / packaging 10 | .Python 11 | build/ 12 | develop-eggs/ 13 | dist/ 14 | downloads/ 15 | eggs/ 16 | .eggs/ 17 | lib/ 18 | lib64/ 19 | parts/ 20 | sdist/ 21 | var/ 22 | wheels/ 23 | *.egg-info/ 24 | .installed.cfg 25 | *.egg 26 | MANIFEST 27 | 28 | # PyInstaller 29 | # Usually these files are written by a python script from a template 30 | # before PyInstaller builds the exe, so as to inject date/other infos into it. 31 | *.manifest 32 | *.spec 33 | 34 | # Installer logs 35 | pip-log.txt 36 | pip-delete-this-directory.txt 37 | 38 | # Unit test / coverage reports 39 | htmlcov/ 40 | .tox/ 41 | .coverage 42 | .coverage.* 43 | .cache 44 | nosetests.xml 45 | coverage.xml 46 | *.cover 47 | .hypothesis/ 48 | .pytest_cache/ 49 | 50 | # Translations 51 | *.mo 52 | *.pot 53 | 54 | # Django stuff: 55 | *.log 56 | local_settings.py 57 | db.sqlite3 58 | 59 | # Flask stuff: 60 | instance/ 61 | .webassets-cache 62 | 63 | # Scrapy stuff: 64 | .scrapy 65 | 66 | # Sphinx documentation 67 | docs/_build/ 68 | 69 | # PyBuilder 70 | target/ 71 | 72 | # Jupyter Notebook 73 | .ipynb_checkpoints 74 | 75 | # pyenv 76 | .python-version 77 | 78 | # celery beat schedule file 79 | celerybeat-schedule 80 | 81 | # SageMath parsed files 82 | *.sage.py 83 | 84 | # Environments 85 | .env 86 | .venv 87 | env/ 88 | venv/ 89 | ENV/ 90 | env.bak/ 91 | venv.bak/ 92 | 93 | # Spyder project settings 94 | .spyderproject 95 | .spyproject 96 | 97 | # Rope project settings 98 | .ropeproject 99 | 100 | # mkdocs documentation 101 | /site 102 | 103 | # mypy 104 | .mypy_cache/ 105 | 106 | # KDevelop 107 | .kdev4/ 108 | *.kdev4 109 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | GNU GENERAL PUBLIC LICENSE 2 | Version 3, 29 June 2007 3 | 4 | Copyright (C) 2007 Free Software Foundation, Inc. 5 | Everyone is permitted to copy and distribute verbatim copies 6 | of this license document, but changing it is not allowed. 7 | 8 | Preamble 9 | 10 | The GNU General Public License is a free, copyleft license for 11 | software and other kinds of works. 12 | 13 | The licenses for most software and other practical works are designed 14 | to take away your freedom to share and change the works. By contrast, 15 | the GNU General Public License is intended to guarantee your freedom to 16 | share and change all versions of a program--to make sure it remains free 17 | software for all its users. We, the Free Software Foundation, use the 18 | GNU General Public License for most of our software; it applies also to 19 | any other work released this way by its authors. You can apply it to 20 | your programs, too. 21 | 22 | When we speak of free software, we are referring to freedom, not 23 | price. Our General Public Licenses are designed to make sure that you 24 | have the freedom to distribute copies of free software (and charge for 25 | them if you wish), that you receive source code or can get it if you 26 | want it, that you can change the software or use pieces of it in new 27 | free programs, and that you know you can do these things. 28 | 29 | To protect your rights, we need to prevent others from denying you 30 | these rights or asking you to surrender the rights. Therefore, you have 31 | certain responsibilities if you distribute copies of the software, or if 32 | you modify it: responsibilities to respect the freedom of others. 33 | 34 | For example, if you distribute copies of such a program, whether 35 | gratis or for a fee, you must pass on to the recipients the same 36 | freedoms that you received. You must make sure that they, too, receive 37 | or can get the source code. And you must show them these terms so they 38 | know their rights. 39 | 40 | Developers that use the GNU GPL protect your rights with two steps: 41 | (1) assert copyright on the software, and (2) offer you this License 42 | giving you legal permission to copy, distribute and/or modify it. 43 | 44 | For the developers' and authors' protection, the GPL clearly explains 45 | that there is no warranty for this free software. For both users' and 46 | authors' sake, the GPL requires that modified versions be marked as 47 | changed, so that their problems will not be attributed erroneously to 48 | authors of previous versions. 49 | 50 | Some devices are designed to deny users access to install or run 51 | modified versions of the software inside them, although the manufacturer 52 | can do so. This is fundamentally incompatible with the aim of 53 | protecting users' freedom to change the software. The systematic 54 | pattern of such abuse occurs in the area of products for individuals to 55 | use, which is precisely where it is most unacceptable. Therefore, we 56 | have designed this version of the GPL to prohibit the practice for those 57 | products. If such problems arise substantially in other domains, we 58 | stand ready to extend this provision to those domains in future versions 59 | of the GPL, as needed to protect the freedom of users. 60 | 61 | Finally, every program is threatened constantly by software patents. 62 | States should not allow patents to restrict development and use of 63 | software on general-purpose computers, but in those that do, we wish to 64 | avoid the special danger that patents applied to a free program could 65 | make it effectively proprietary. To prevent this, the GPL assures that 66 | patents cannot be used to render the program non-free. 67 | 68 | The precise terms and conditions for copying, distribution and 69 | modification follow. 70 | 71 | TERMS AND CONDITIONS 72 | 73 | 0. Definitions. 74 | 75 | "This License" refers to version 3 of the GNU General Public License. 76 | 77 | "Copyright" also means copyright-like laws that apply to other kinds of 78 | works, such as semiconductor masks. 79 | 80 | "The Program" refers to any copyrightable work licensed under this 81 | License. Each licensee is addressed as "you". "Licensees" and 82 | "recipients" may be individuals or organizations. 83 | 84 | To "modify" a work means to copy from or adapt all or part of the work 85 | in a fashion requiring copyright permission, other than the making of an 86 | exact copy. The resulting work is called a "modified version" of the 87 | earlier work or a work "based on" the earlier work. 88 | 89 | A "covered work" means either the unmodified Program or a work based 90 | on the Program. 91 | 92 | To "propagate" a work means to do anything with it that, without 93 | permission, would make you directly or secondarily liable for 94 | infringement under applicable copyright law, except executing it on a 95 | computer or modifying a private copy. Propagation includes copying, 96 | distribution (with or without modification), making available to the 97 | public, and in some countries other activities as well. 98 | 99 | To "convey" a work means any kind of propagation that enables other 100 | parties to make or receive copies. Mere interaction with a user through 101 | a computer network, with no transfer of a copy, is not conveying. 102 | 103 | An interactive user interface displays "Appropriate Legal Notices" 104 | to the extent that it includes a convenient and prominently visible 105 | feature that (1) displays an appropriate copyright notice, and (2) 106 | tells the user that there is no warranty for the work (except to the 107 | extent that warranties are provided), that licensees may convey the 108 | work under this License, and how to view a copy of this License. If 109 | the interface presents a list of user commands or options, such as a 110 | menu, a prominent item in the list meets this criterion. 111 | 112 | 1. Source Code. 113 | 114 | The "source code" for a work means the preferred form of the work 115 | for making modifications to it. "Object code" means any non-source 116 | form of a work. 117 | 118 | A "Standard Interface" means an interface that either is an official 119 | standard defined by a recognized standards body, or, in the case of 120 | interfaces specified for a particular programming language, one that 121 | is widely used among developers working in that language. 122 | 123 | The "System Libraries" of an executable work include anything, other 124 | than the work as a whole, that (a) is included in the normal form of 125 | packaging a Major Component, but which is not part of that Major 126 | Component, and (b) serves only to enable use of the work with that 127 | Major Component, or to implement a Standard Interface for which an 128 | implementation is available to the public in source code form. A 129 | "Major Component", in this context, means a major essential component 130 | (kernel, window system, and so on) of the specific operating system 131 | (if any) on which the executable work runs, or a compiler used to 132 | produce the work, or an object code interpreter used to run it. 133 | 134 | The "Corresponding Source" for a work in object code form means all 135 | the source code needed to generate, install, and (for an executable 136 | work) run the object code and to modify the work, including scripts to 137 | control those activities. However, it does not include the work's 138 | System Libraries, or general-purpose tools or generally available free 139 | programs which are used unmodified in performing those activities but 140 | which are not part of the work. For example, Corresponding Source 141 | includes interface definition files associated with source files for 142 | the work, and the source code for shared libraries and dynamically 143 | linked subprograms that the work is specifically designed to require, 144 | such as by intimate data communication or control flow between those 145 | subprograms and other parts of the work. 146 | 147 | The Corresponding Source need not include anything that users 148 | can regenerate automatically from other parts of the Corresponding 149 | Source. 150 | 151 | The Corresponding Source for a work in source code form is that 152 | same work. 153 | 154 | 2. Basic Permissions. 155 | 156 | All rights granted under this License are granted for the term of 157 | copyright on the Program, and are irrevocable provided the stated 158 | conditions are met. This License explicitly affirms your unlimited 159 | permission to run the unmodified Program. The output from running a 160 | covered work is covered by this License only if the output, given its 161 | content, constitutes a covered work. This License acknowledges your 162 | rights of fair use or other equivalent, as provided by copyright law. 163 | 164 | You may make, run and propagate covered works that you do not 165 | convey, without conditions so long as your license otherwise remains 166 | in force. You may convey covered works to others for the sole purpose 167 | of having them make modifications exclusively for you, or provide you 168 | with facilities for running those works, provided that you comply with 169 | the terms of this License in conveying all material for which you do 170 | not control copyright. Those thus making or running the covered works 171 | for you must do so exclusively on your behalf, under your direction 172 | and control, on terms that prohibit them from making any copies of 173 | your copyrighted material outside their relationship with you. 174 | 175 | Conveying under any other circumstances is permitted solely under 176 | the conditions stated below. Sublicensing is not allowed; section 10 177 | makes it unnecessary. 178 | 179 | 3. Protecting Users' Legal Rights From Anti-Circumvention Law. 180 | 181 | No covered work shall be deemed part of an effective technological 182 | measure under any applicable law fulfilling obligations under article 183 | 11 of the WIPO copyright treaty adopted on 20 December 1996, or 184 | similar laws prohibiting or restricting circumvention of such 185 | measures. 186 | 187 | When you convey a covered work, you waive any legal power to forbid 188 | circumvention of technological measures to the extent such circumvention 189 | is effected by exercising rights under this License with respect to 190 | the covered work, and you disclaim any intention to limit operation or 191 | modification of the work as a means of enforcing, against the work's 192 | users, your or third parties' legal rights to forbid circumvention of 193 | technological measures. 194 | 195 | 4. Conveying Verbatim Copies. 196 | 197 | You may convey verbatim copies of the Program's source code as you 198 | receive it, in any medium, provided that you conspicuously and 199 | appropriately publish on each copy an appropriate copyright notice; 200 | keep intact all notices stating that this License and any 201 | non-permissive terms added in accord with section 7 apply to the code; 202 | keep intact all notices of the absence of any warranty; and give all 203 | recipients a copy of this License along with the Program. 204 | 205 | You may charge any price or no price for each copy that you convey, 206 | and you may offer support or warranty protection for a fee. 207 | 208 | 5. Conveying Modified Source Versions. 209 | 210 | You may convey a work based on the Program, or the modifications to 211 | produce it from the Program, in the form of source code under the 212 | terms of section 4, provided that you also meet all of these conditions: 213 | 214 | a) The work must carry prominent notices stating that you modified 215 | it, and giving a relevant date. 216 | 217 | b) The work must carry prominent notices stating that it is 218 | released under this License and any conditions added under section 219 | 7. This requirement modifies the requirement in section 4 to 220 | "keep intact all notices". 221 | 222 | c) You must license the entire work, as a whole, under this 223 | License to anyone who comes into possession of a copy. This 224 | License will therefore apply, along with any applicable section 7 225 | additional terms, to the whole of the work, and all its parts, 226 | regardless of how they are packaged. This License gives no 227 | permission to license the work in any other way, but it does not 228 | invalidate such permission if you have separately received it. 229 | 230 | d) If the work has interactive user interfaces, each must display 231 | Appropriate Legal Notices; however, if the Program has interactive 232 | interfaces that do not display Appropriate Legal Notices, your 233 | work need not make them do so. 234 | 235 | A compilation of a covered work with other separate and independent 236 | works, which are not by their nature extensions of the covered work, 237 | and which are not combined with it such as to form a larger program, 238 | in or on a volume of a storage or distribution medium, is called an 239 | "aggregate" if the compilation and its resulting copyright are not 240 | used to limit the access or legal rights of the compilation's users 241 | beyond what the individual works permit. Inclusion of a covered work 242 | in an aggregate does not cause this License to apply to the other 243 | parts of the aggregate. 244 | 245 | 6. Conveying Non-Source Forms. 246 | 247 | You may convey a covered work in object code form under the terms 248 | of sections 4 and 5, provided that you also convey the 249 | machine-readable Corresponding Source under the terms of this License, 250 | in one of these ways: 251 | 252 | a) Convey the object code in, or embodied in, a physical product 253 | (including a physical distribution medium), accompanied by the 254 | Corresponding Source fixed on a durable physical medium 255 | customarily used for software interchange. 256 | 257 | b) Convey the object code in, or embodied in, a physical product 258 | (including a physical distribution medium), accompanied by a 259 | written offer, valid for at least three years and valid for as 260 | long as you offer spare parts or customer support for that product 261 | model, to give anyone who possesses the object code either (1) a 262 | copy of the Corresponding Source for all the software in the 263 | product that is covered by this License, on a durable physical 264 | medium customarily used for software interchange, for a price no 265 | more than your reasonable cost of physically performing this 266 | conveying of source, or (2) access to copy the 267 | Corresponding Source from a network server at no charge. 268 | 269 | c) Convey individual copies of the object code with a copy of the 270 | written offer to provide the Corresponding Source. This 271 | alternative is allowed only occasionally and noncommercially, and 272 | only if you received the object code with such an offer, in accord 273 | with subsection 6b. 274 | 275 | d) Convey the object code by offering access from a designated 276 | place (gratis or for a charge), and offer equivalent access to the 277 | Corresponding Source in the same way through the same place at no 278 | further charge. You need not require recipients to copy the 279 | Corresponding Source along with the object code. If the place to 280 | copy the object code is a network server, the Corresponding Source 281 | may be on a different server (operated by you or a third party) 282 | that supports equivalent copying facilities, provided you maintain 283 | clear directions next to the object code saying where to find the 284 | Corresponding Source. Regardless of what server hosts the 285 | Corresponding Source, you remain obligated to ensure that it is 286 | available for as long as needed to satisfy these requirements. 287 | 288 | e) Convey the object code using peer-to-peer transmission, provided 289 | you inform other peers where the object code and Corresponding 290 | Source of the work are being offered to the general public at no 291 | charge under subsection 6d. 292 | 293 | A separable portion of the object code, whose source code is excluded 294 | from the Corresponding Source as a System Library, need not be 295 | included in conveying the object code work. 296 | 297 | A "User Product" is either (1) a "consumer product", which means any 298 | tangible personal property which is normally used for personal, family, 299 | or household purposes, or (2) anything designed or sold for incorporation 300 | into a dwelling. In determining whether a product is a consumer product, 301 | doubtful cases shall be resolved in favor of coverage. For a particular 302 | product received by a particular user, "normally used" refers to a 303 | typical or common use of that class of product, regardless of the status 304 | of the particular user or of the way in which the particular user 305 | actually uses, or expects or is expected to use, the product. A product 306 | is a consumer product regardless of whether the product has substantial 307 | commercial, industrial or non-consumer uses, unless such uses represent 308 | the only significant mode of use of the product. 309 | 310 | "Installation Information" for a User Product means any methods, 311 | procedures, authorization keys, or other information required to install 312 | and execute modified versions of a covered work in that User Product from 313 | a modified version of its Corresponding Source. The information must 314 | suffice to ensure that the continued functioning of the modified object 315 | code is in no case prevented or interfered with solely because 316 | modification has been made. 317 | 318 | If you convey an object code work under this section in, or with, or 319 | specifically for use in, a User Product, and the conveying occurs as 320 | part of a transaction in which the right of possession and use of the 321 | User Product is transferred to the recipient in perpetuity or for a 322 | fixed term (regardless of how the transaction is characterized), the 323 | Corresponding Source conveyed under this section must be accompanied 324 | by the Installation Information. But this requirement does not apply 325 | if neither you nor any third party retains the ability to install 326 | modified object code on the User Product (for example, the work has 327 | been installed in ROM). 328 | 329 | The requirement to provide Installation Information does not include a 330 | requirement to continue to provide support service, warranty, or updates 331 | for a work that has been modified or installed by the recipient, or for 332 | the User Product in which it has been modified or installed. Access to a 333 | network may be denied when the modification itself materially and 334 | adversely affects the operation of the network or violates the rules and 335 | protocols for communication across the network. 336 | 337 | Corresponding Source conveyed, and Installation Information provided, 338 | in accord with this section must be in a format that is publicly 339 | documented (and with an implementation available to the public in 340 | source code form), and must require no special password or key for 341 | unpacking, reading or copying. 342 | 343 | 7. Additional Terms. 344 | 345 | "Additional permissions" are terms that supplement the terms of this 346 | License by making exceptions from one or more of its conditions. 347 | Additional permissions that are applicable to the entire Program shall 348 | be treated as though they were included in this License, to the extent 349 | that they are valid under applicable law. If additional permissions 350 | apply only to part of the Program, that part may be used separately 351 | under those permissions, but the entire Program remains governed by 352 | this License without regard to the additional permissions. 353 | 354 | When you convey a copy of a covered work, you may at your option 355 | remove any additional permissions from that copy, or from any part of 356 | it. (Additional permissions may be written to require their own 357 | removal in certain cases when you modify the work.) You may place 358 | additional permissions on material, added by you to a covered work, 359 | for which you have or can give appropriate copyright permission. 360 | 361 | Notwithstanding any other provision of this License, for material you 362 | add to a covered work, you may (if authorized by the copyright holders of 363 | that material) supplement the terms of this License with terms: 364 | 365 | a) Disclaiming warranty or limiting liability differently from the 366 | terms of sections 15 and 16 of this License; or 367 | 368 | b) Requiring preservation of specified reasonable legal notices or 369 | author attributions in that material or in the Appropriate Legal 370 | Notices displayed by works containing it; or 371 | 372 | c) Prohibiting misrepresentation of the origin of that material, or 373 | requiring that modified versions of such material be marked in 374 | reasonable ways as different from the original version; or 375 | 376 | d) Limiting the use for publicity purposes of names of licensors or 377 | authors of the material; or 378 | 379 | e) Declining to grant rights under trademark law for use of some 380 | trade names, trademarks, or service marks; or 381 | 382 | f) Requiring indemnification of licensors and authors of that 383 | material by anyone who conveys the material (or modified versions of 384 | it) with contractual assumptions of liability to the recipient, for 385 | any liability that these contractual assumptions directly impose on 386 | those licensors and authors. 387 | 388 | All other non-permissive additional terms are considered "further 389 | restrictions" within the meaning of section 10. If the Program as you 390 | received it, or any part of it, contains a notice stating that it is 391 | governed by this License along with a term that is a further 392 | restriction, you may remove that term. If a license document contains 393 | a further restriction but permits relicensing or conveying under this 394 | License, you may add to a covered work material governed by the terms 395 | of that license document, provided that the further restriction does 396 | not survive such relicensing or conveying. 397 | 398 | If you add terms to a covered work in accord with this section, you 399 | must place, in the relevant source files, a statement of the 400 | additional terms that apply to those files, or a notice indicating 401 | where to find the applicable terms. 402 | 403 | Additional terms, permissive or non-permissive, may be stated in the 404 | form of a separately written license, or stated as exceptions; 405 | the above requirements apply either way. 406 | 407 | 8. Termination. 408 | 409 | You may not propagate or modify a covered work except as expressly 410 | provided under this License. Any attempt otherwise to propagate or 411 | modify it is void, and will automatically terminate your rights under 412 | this License (including any patent licenses granted under the third 413 | paragraph of section 11). 414 | 415 | However, if you cease all violation of this License, then your 416 | license from a particular copyright holder is reinstated (a) 417 | provisionally, unless and until the copyright holder explicitly and 418 | finally terminates your license, and (b) permanently, if the copyright 419 | holder fails to notify you of the violation by some reasonable means 420 | prior to 60 days after the cessation. 421 | 422 | Moreover, your license from a particular copyright holder is 423 | reinstated permanently if the copyright holder notifies you of the 424 | violation by some reasonable means, this is the first time you have 425 | received notice of violation of this License (for any work) from that 426 | copyright holder, and you cure the violation prior to 30 days after 427 | your receipt of the notice. 428 | 429 | Termination of your rights under this section does not terminate the 430 | licenses of parties who have received copies or rights from you under 431 | this License. If your rights have been terminated and not permanently 432 | reinstated, you do not qualify to receive new licenses for the same 433 | material under section 10. 434 | 435 | 9. Acceptance Not Required for Having Copies. 436 | 437 | You are not required to accept this License in order to receive or 438 | run a copy of the Program. Ancillary propagation of a covered work 439 | occurring solely as a consequence of using peer-to-peer transmission 440 | to receive a copy likewise does not require acceptance. However, 441 | nothing other than this License grants you permission to propagate or 442 | modify any covered work. These actions infringe copyright if you do 443 | not accept this License. Therefore, by modifying or propagating a 444 | covered work, you indicate your acceptance of this License to do so. 445 | 446 | 10. Automatic Licensing of Downstream Recipients. 447 | 448 | Each time you convey a covered work, the recipient automatically 449 | receives a license from the original licensors, to run, modify and 450 | propagate that work, subject to this License. You are not responsible 451 | for enforcing compliance by third parties with this License. 452 | 453 | An "entity transaction" is a transaction transferring control of an 454 | organization, or substantially all assets of one, or subdividing an 455 | organization, or merging organizations. If propagation of a covered 456 | work results from an entity transaction, each party to that 457 | transaction who receives a copy of the work also receives whatever 458 | licenses to the work the party's predecessor in interest had or could 459 | give under the previous paragraph, plus a right to possession of the 460 | Corresponding Source of the work from the predecessor in interest, if 461 | the predecessor has it or can get it with reasonable efforts. 462 | 463 | You may not impose any further restrictions on the exercise of the 464 | rights granted or affirmed under this License. For example, you may 465 | not impose a license fee, royalty, or other charge for exercise of 466 | rights granted under this License, and you may not initiate litigation 467 | (including a cross-claim or counterclaim in a lawsuit) alleging that 468 | any patent claim is infringed by making, using, selling, offering for 469 | sale, or importing the Program or any portion of it. 470 | 471 | 11. Patents. 472 | 473 | A "contributor" is a copyright holder who authorizes use under this 474 | License of the Program or a work on which the Program is based. The 475 | work thus licensed is called the contributor's "contributor version". 476 | 477 | A contributor's "essential patent claims" are all patent claims 478 | owned or controlled by the contributor, whether already acquired or 479 | hereafter acquired, that would be infringed by some manner, permitted 480 | by this License, of making, using, or selling its contributor version, 481 | but do not include claims that would be infringed only as a 482 | consequence of further modification of the contributor version. For 483 | purposes of this definition, "control" includes the right to grant 484 | patent sublicenses in a manner consistent with the requirements of 485 | this License. 486 | 487 | Each contributor grants you a non-exclusive, worldwide, royalty-free 488 | patent license under the contributor's essential patent claims, to 489 | make, use, sell, offer for sale, import and otherwise run, modify and 490 | propagate the contents of its contributor version. 491 | 492 | In the following three paragraphs, a "patent license" is any express 493 | agreement or commitment, however denominated, not to enforce a patent 494 | (such as an express permission to practice a patent or covenant not to 495 | sue for patent infringement). To "grant" such a patent license to a 496 | party means to make such an agreement or commitment not to enforce a 497 | patent against the party. 498 | 499 | If you convey a covered work, knowingly relying on a patent license, 500 | and the Corresponding Source of the work is not available for anyone 501 | to copy, free of charge and under the terms of this License, through a 502 | publicly available network server or other readily accessible means, 503 | then you must either (1) cause the Corresponding Source to be so 504 | available, or (2) arrange to deprive yourself of the benefit of the 505 | patent license for this particular work, or (3) arrange, in a manner 506 | consistent with the requirements of this License, to extend the patent 507 | license to downstream recipients. "Knowingly relying" means you have 508 | actual knowledge that, but for the patent license, your conveying the 509 | covered work in a country, or your recipient's use of the covered work 510 | in a country, would infringe one or more identifiable patents in that 511 | country that you have reason to believe are valid. 512 | 513 | If, pursuant to or in connection with a single transaction or 514 | arrangement, you convey, or propagate by procuring conveyance of, a 515 | covered work, and grant a patent license to some of the parties 516 | receiving the covered work authorizing them to use, propagate, modify 517 | or convey a specific copy of the covered work, then the patent license 518 | you grant is automatically extended to all recipients of the covered 519 | work and works based on it. 520 | 521 | A patent license is "discriminatory" if it does not include within 522 | the scope of its coverage, prohibits the exercise of, or is 523 | conditioned on the non-exercise of one or more of the rights that are 524 | specifically granted under this License. You may not convey a covered 525 | work if you are a party to an arrangement with a third party that is 526 | in the business of distributing software, under which you make payment 527 | to the third party based on the extent of your activity of conveying 528 | the work, and under which the third party grants, to any of the 529 | parties who would receive the covered work from you, a discriminatory 530 | patent license (a) in connection with copies of the covered work 531 | conveyed by you (or copies made from those copies), or (b) primarily 532 | for and in connection with specific products or compilations that 533 | contain the covered work, unless you entered into that arrangement, 534 | or that patent license was granted, prior to 28 March 2007. 535 | 536 | Nothing in this License shall be construed as excluding or limiting 537 | any implied license or other defenses to infringement that may 538 | otherwise be available to you under applicable patent law. 539 | 540 | 12. No Surrender of Others' Freedom. 541 | 542 | If conditions are imposed on you (whether by court order, agreement or 543 | otherwise) that contradict the conditions of this License, they do not 544 | excuse you from the conditions of this License. If you cannot convey a 545 | covered work so as to satisfy simultaneously your obligations under this 546 | License and any other pertinent obligations, then as a consequence you may 547 | not convey it at all. For example, if you agree to terms that obligate you 548 | to collect a royalty for further conveying from those to whom you convey 549 | the Program, the only way you could satisfy both those terms and this 550 | License would be to refrain entirely from conveying the Program. 551 | 552 | 13. Use with the GNU Affero General Public License. 553 | 554 | Notwithstanding any other provision of this License, you have 555 | permission to link or combine any covered work with a work licensed 556 | under version 3 of the GNU Affero General Public License into a single 557 | combined work, and to convey the resulting work. The terms of this 558 | License will continue to apply to the part which is the covered work, 559 | but the special requirements of the GNU Affero General Public License, 560 | section 13, concerning interaction through a network will apply to the 561 | combination as such. 562 | 563 | 14. Revised Versions of this License. 564 | 565 | The Free Software Foundation may publish revised and/or new versions of 566 | the GNU General Public License from time to time. Such new versions will 567 | be similar in spirit to the present version, but may differ in detail to 568 | address new problems or concerns. 569 | 570 | Each version is given a distinguishing version number. If the 571 | Program specifies that a certain numbered version of the GNU General 572 | Public License "or any later version" applies to it, you have the 573 | option of following the terms and conditions either of that numbered 574 | version or of any later version published by the Free Software 575 | Foundation. If the Program does not specify a version number of the 576 | GNU General Public License, you may choose any version ever published 577 | by the Free Software Foundation. 578 | 579 | If the Program specifies that a proxy can decide which future 580 | versions of the GNU General Public License can be used, that proxy's 581 | public statement of acceptance of a version permanently authorizes you 582 | to choose that version for the Program. 583 | 584 | Later license versions may give you additional or different 585 | permissions. However, no additional obligations are imposed on any 586 | author or copyright holder as a result of your choosing to follow a 587 | later version. 588 | 589 | 15. Disclaimer of Warranty. 590 | 591 | THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY 592 | APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT 593 | HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY 594 | OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, 595 | THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 596 | PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM 597 | IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF 598 | ALL NECESSARY SERVICING, REPAIR OR CORRECTION. 599 | 600 | 16. Limitation of Liability. 601 | 602 | IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING 603 | WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS 604 | THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY 605 | GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE 606 | USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF 607 | DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD 608 | PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), 609 | EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF 610 | SUCH DAMAGES. 611 | 612 | 17. Interpretation of Sections 15 and 16. 613 | 614 | If the disclaimer of warranty and limitation of liability provided 615 | above cannot be given local legal effect according to their terms, 616 | reviewing courts shall apply local law that most closely approximates 617 | an absolute waiver of all civil liability in connection with the 618 | Program, unless a warranty or assumption of liability accompanies a 619 | copy of the Program in return for a fee. 620 | 621 | END OF TERMS AND CONDITIONS 622 | 623 | How to Apply These Terms to Your New Programs 624 | 625 | If you develop a new program, and you want it to be of the greatest 626 | possible use to the public, the best way to achieve this is to make it 627 | free software which everyone can redistribute and change under these terms. 628 | 629 | To do so, attach the following notices to the program. It is safest 630 | to attach them to the start of each source file to most effectively 631 | state the exclusion of warranty; and each file should have at least 632 | the "copyright" line and a pointer to where the full notice is found. 633 | 634 | 635 | Copyright (C) 636 | 637 | This program is free software: you can redistribute it and/or modify 638 | it under the terms of the GNU General Public License as published by 639 | the Free Software Foundation, either version 3 of the License, or 640 | (at your option) any later version. 641 | 642 | This program is distributed in the hope that it will be useful, 643 | but WITHOUT ANY WARRANTY; without even the implied warranty of 644 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 645 | GNU General Public License for more details. 646 | 647 | You should have received a copy of the GNU General Public License 648 | along with this program. If not, see . 649 | 650 | Also add information on how to contact you by electronic and paper mail. 651 | 652 | If the program does terminal interaction, make it output a short 653 | notice like this when it starts in an interactive mode: 654 | 655 | Copyright (C) 656 | This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'. 657 | This is free software, and you are welcome to redistribute it 658 | under certain conditions; type `show c' for details. 659 | 660 | The hypothetical commands `show w' and `show c' should show the appropriate 661 | parts of the General Public License. Of course, your program's commands 662 | might be different; for a GUI interface, you would use an "about box". 663 | 664 | You should also get your employer (if you work as a programmer) or school, 665 | if any, to sign a "copyright disclaimer" for the program, if necessary. 666 | For more information on this, and how to apply and follow the GNU GPL, see 667 | . 668 | 669 | The GNU General Public License does not permit incorporating your program 670 | into proprietary programs. If your program is a subroutine library, you 671 | may consider it more useful to permit linking proprietary applications with 672 | the library. If this is what you want to do, use the GNU Lesser General 673 | Public License instead of this License. But first, please read 674 | . 675 | -------------------------------------------------------------------------------- /README.rst: -------------------------------------------------------------------------------- 1 | ===== 2 | Umbra 3 | ===== 4 | 5 | MOVED TO GITLAB 6 | =============== 7 | 8 | POP projects developed by Saltstack are being moved to Gitlab. 9 | 10 | The new location of idem is here: 11 | 12 | https://gitlab.com/saltstack/pop/umbra 13 | -------------------------------------------------------------------------------- /demo/mkdata.py: -------------------------------------------------------------------------------- 1 | import faker 2 | import random 3 | import json 4 | import pprint 5 | import os 6 | import sys 7 | 8 | USERS = ['thatch', 'frank', 'bob', 'sud', 'mary'] 9 | IDS = ['ragnarok', 'thor', 'odin', 'loki'] 10 | 11 | 12 | def gen_events(dates, cmds): 13 | ret = [] 14 | for cmd in cmds: 15 | id_ = random.choice(IDS) 16 | pid = random.randint(1024, 2**16) 17 | event = { 18 | 'tag': 'salt/beacon/{}/sh/{}'.format(id_, pid), 19 | 'data': {'_stamp': next(dates)[0].isoformat(), 'cmd': cmd['cmd'], 'user': random.choice(USERS), 'args': cmd['args'], 'id': id_}, 20 | } 21 | ret.append(event) 22 | return ret 23 | 24 | 25 | def mkdata(size=500000): 26 | fake = faker.Faker() 27 | 28 | dates = fake.time_series('-1000d', precision=1) 29 | 30 | ret = [] 31 | cmds = [] 32 | icmds = [] 33 | with open(os.path.join(os.path.expanduser('~'), '.bash_history'), 'r') as rfp: 34 | pcmds = rfp.readlines() 35 | for cmd in pcmds: 36 | if ';' in cmd: 37 | icmds.extend(cmd.split(';')) 38 | elif '||' in cmd: 39 | icmds.extend(cmd.split('||')) 40 | elif '&&' in cmd: 41 | icmds.extend(cmd.split('&&')) 42 | else: 43 | icmds.append(cmd) 44 | for cmd in icmds: 45 | parts = cmd.split() 46 | final = {'cmd': parts[0]} 47 | if len(parts) > 1: 48 | final['args'] = parts[1:] 49 | else: 50 | final['args'] = [] 51 | cmds.append(final) 52 | while len(ret) < size: 53 | ret.extend(gen_events(dates, cmds)) 54 | return ret 55 | 56 | def save(): 57 | if len(sys.argv) > 1: 58 | size = int(sys.argv[1]) 59 | else: 60 | size = 500000 61 | ret = mkdata(size) 62 | #pprint.pprint(ret) 63 | with open('shell.json', 'w+') as wfp: 64 | wfp.write(json.dumps(ret)) 65 | 66 | 67 | save() 68 | -------------------------------------------------------------------------------- /demo/reactor/allproc.sls: -------------------------------------------------------------------------------- 1 | kill_rogue: 2 | local.ps.pkill: 3 | - tgt: {{ data['id'] }} 4 | - tgt_type: glob 5 | - args: 6 | - pattern: {{ data['name'] }} 7 | -------------------------------------------------------------------------------- /demo/reactor/event.yml: -------------------------------------------------------------------------------- 1 | allproc: 2 | ingress: 3 | salt_event: 'salt/beacon/*/allproc*' 4 | data: salt_event 5 | model: hbos 6 | egress: 7 | - cli 8 | - salt_event 9 | train_for: 10000 10 | enabled: 1 11 | -------------------------------------------------------------------------------- /demo/reactor/master: -------------------------------------------------------------------------------- 1 | reactor: 2 | - 'umbra/allproc*': 3 | - /srv/reactor/allproc.sls -------------------------------------------------------------------------------- /demo/salt/beacons/allproc.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | ''' 3 | Send events covering process status 4 | ''' 5 | 6 | # Import Python Libs 7 | from __future__ import absolute_import, unicode_literals 8 | import logging 9 | 10 | # Import third party libs 11 | # pylint: disable=import-error 12 | try: 13 | import salt.utils.psutil_compat as psutil 14 | HAS_PSUTIL = True 15 | except ImportError: 16 | HAS_PSUTIL = False 17 | 18 | __virtualname__ = 'allproc' 19 | 20 | 21 | def __virtual__(): 22 | if not HAS_PSUTIL: 23 | return (False, 'cannot load ps beacon: psutil not available') 24 | return __virtualname__ 25 | 26 | 27 | def validate(config): 28 | return True, 'No configuration yet required' 29 | 30 | 31 | def beacon(config): 32 | ''' 33 | Scan for processes and fire events with all process data 34 | 35 | Example Config 36 | 37 | .. code-block:: yaml 38 | 39 | beacons: 40 | allproc: [] 41 | ''' 42 | ret = [] 43 | procs = set() 44 | for proc in psutil.process_iter(): 45 | name = proc.name() 46 | # These rotating kworkers polute the dataset 47 | if name.startswith('kworker/'): 48 | continue 49 | procs.add(name) 50 | for name in procs: 51 | ret.append({'name': name}) 52 | return ret 53 | 54 | -------------------------------------------------------------------------------- /docs/Makefile: -------------------------------------------------------------------------------- 1 | # Minimal makefile for Sphinx documentation 2 | # 3 | 4 | # You can set these variables from the command line. 5 | SPHINXOPTS = 6 | SPHINXBUILD = sphinx-build 7 | SOURCEDIR = . 8 | BUILDDIR = _build 9 | 10 | # Put it first so that "make" without argument is like "make help". 11 | help: 12 | @$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O) 13 | 14 | .PHONY: help Makefile 15 | 16 | # Catch-all target: route all unknown targets to Sphinx using the new 17 | # "make mode" option. $(O) is meant as a shortcut for $(SPHINXOPTS). 18 | %: Makefile 19 | @$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O) -------------------------------------------------------------------------------- /docs/conf.py: -------------------------------------------------------------------------------- 1 | # Configuration file for the Sphinx documentation builder. 2 | # 3 | # This file only contains a selection of the most common options. For a full 4 | # list see the documentation: 5 | # http://www.sphinx-doc.org/en/master/config 6 | 7 | # -- Path setup -------------------------------------------------------------- 8 | 9 | # If extensions (or modules to document with autodoc) are in another directory, 10 | # add these directories to sys.path here. If the directory is relative to the 11 | # documentation root, use os.path.abspath to make it absolute, like shown here. 12 | # 13 | # import os 14 | # import sys 15 | # sys.path.insert(0, os.path.abspath('.')) 16 | 17 | 18 | # -- Project information ----------------------------------------------------- 19 | 20 | project = 'Umbra' 21 | copyright = '2019, Thomas S Hatch' 22 | author = 'Thomas S Hatch' 23 | 24 | # The full version, including alpha/beta/rc tags 25 | release = '1.3.0' 26 | 27 | 28 | # -- General configuration --------------------------------------------------- 29 | 30 | # Add any Sphinx extension module names here, as strings. They can be 31 | # extensions coming with Sphinx (named 'sphinx.ext.*') or your custom 32 | # ones. 33 | extensions = [ 34 | ] 35 | 36 | # Add any paths that contain templates here, relative to this directory. 37 | templates_path = ['_templates'] 38 | 39 | # List of patterns, relative to source directory, that match files and 40 | # directories to ignore when looking for source files. 41 | # This pattern also affects html_static_path and html_extra_path. 42 | exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store'] 43 | 44 | # Master doc 45 | master_doc = 'index' 46 | 47 | # -- Options for HTML output ------------------------------------------------- 48 | 49 | # The theme to use for HTML and HTML Help pages. See the documentation for 50 | # a list of builtin themes. 51 | # 52 | html_theme = 'alabaster' 53 | 54 | # Add any paths that contain custom static files (such as style sheets) here, 55 | # relative to this directory. They are copied after the builtin static files, 56 | # so a file named "default.css" will overwrite the builtin "default.css". 57 | html_static_path = ['_static'] 58 | -------------------------------------------------------------------------------- /docs/index.rst: -------------------------------------------------------------------------------- 1 | .. Umbra documentation master file, created by 2 | sphinx-quickstart on Tue May 21 10:31:31 2019. 3 | You can adapt this file completely to your liking, but it should at least 4 | contain the root `toctree` directive. 5 | 6 | Welcome to Umbra's documentation! 7 | ================================= 8 | 9 | .. toctree:: 10 | :maxdepth: 2 11 | :glob: 12 | 13 | topics/quickstart 14 | topics/salt 15 | topics/flows 16 | topics/ingress 17 | topics/data 18 | topics/models 19 | topics/egress 20 | 21 | Indices and tables 22 | ================== 23 | 24 | * :ref:`genindex` 25 | * :ref:`modindex` 26 | * :ref:`search` 27 | -------------------------------------------------------------------------------- /docs/make.bat: -------------------------------------------------------------------------------- 1 | @ECHO OFF 2 | 3 | pushd %~dp0 4 | 5 | REM Command file for Sphinx documentation 6 | 7 | if "%SPHINXBUILD%" == "" ( 8 | set SPHINXBUILD=sphinx-build 9 | ) 10 | set SOURCEDIR=. 11 | set BUILDDIR=_build 12 | 13 | if "%1" == "" goto help 14 | 15 | %SPHINXBUILD% >NUL 2>NUL 16 | if errorlevel 9009 ( 17 | echo. 18 | echo.The 'sphinx-build' command was not found. Make sure you have Sphinx 19 | echo.installed, then set the SPHINXBUILD environment variable to point 20 | echo.to the full path of the 'sphinx-build' executable. Alternatively you 21 | echo.may add the Sphinx directory to PATH. 22 | echo. 23 | echo.If you don't have Sphinx installed, grab it from 24 | echo.http://sphinx-doc.org/ 25 | exit /b 1 26 | ) 27 | 28 | %SPHINXBUILD% -M %1 %SOURCEDIR% %BUILDDIR% %SPHINXOPTS% 29 | goto end 30 | 31 | :help 32 | %SPHINXBUILD% -M help %SOURCEDIR% %BUILDDIR% %SPHINXOPTS% 33 | 34 | :end 35 | popd 36 | -------------------------------------------------------------------------------- /docs/topics/data.rst: -------------------------------------------------------------------------------- 1 | ==================== 2 | Writing Data Plugins 3 | ==================== 4 | 5 | Data plugins are simple, but they do follow a format and expect certain 6 | inputs and outputs. The data functions are executed as new data comes in 7 | and will be executed over and over again. Unlike, for instance, `ingress` 8 | plugins that are only executed once to set up the flow of data. 9 | 10 | The data plugin takes 2 functions, both are coroutines. The first function, 11 | `prepare` takes the name of the pipe and the inbound data as a list. The 12 | `prepare` coroutine function needs to return a list of arrays that can be 13 | used by a model. 14 | 15 | Prepare Function 16 | ================ 17 | 18 | This example shows how to prepare data formatted in a python dict for 19 | the salt event bus. Remember that Umbra uses `pop` as the plugin system so 20 | the `hub` needs to be accepted as the first argument. 21 | 22 | .. code-block:: python 23 | 24 | async def prepare(hub, pipe, inbound): 25 | ''' 26 | Takes the raw data loaded from a salt event stream, this data is transformed 27 | into data that can be loaded in to the model and we save the string map dicts 28 | for future translation 29 | ''' 30 | ret = [] 31 | d_count = 0 32 | for event in inbound: 33 | if not hub.P[pipe]['tmap_populated']: 34 | for key in event['data']: 35 | if not isinstance(event['data'][key], str): 36 | # TODO: Make this able to handle more than just strings 37 | continue 38 | if key == '_stamp': 39 | continue 40 | hub.P[pipe]['tmap'].append(key) 41 | hub.P[pipe]['tmap_populated'] = True 42 | x = [0 for n in hub.P[pipe]['tmap']] 43 | for key in event['data']: 44 | if not isinstance(event['data'][key], str): 45 | # TODO: Be able to handle more than just strings here 46 | continue 47 | if key == '_stamp': 48 | continue 49 | x[hub.P[pipe]['tmap'].index(key)] = hub.data.init.word_map( 50 | event['data'][key], 51 | hub.P[pipe]['words'], 52 | hub.P[pipe]['r_words']) 53 | ret.append(x) 54 | return ret 55 | 56 | Since the data preparations need to persist data across runs, this is one of the places 57 | in umbra that we need to use the `hub` which has been given to us by `pop`. 58 | 59 | The `hub` is a namespaced hierarchy used to store plugin references and variables. 60 | The `hub` makes it easy to persist data in a clean way across the entire application. 61 | In this case we use the `hub.P` dict that has already been prepared for you. Under 62 | hub.P there is a dict for the pipe we are running in, and this is the place to store 63 | data about this pipe. 64 | 65 | With the `hub` already prepared we can store information that we need in future runs, like 66 | the words and reverse words dicts and we use the tmap to line up the correct keys from the 67 | events in the correct locations in the returned number array. 68 | 69 | Refine Function 70 | =============== 71 | 72 | After creating the `prepare` function the `refine` coroutine is also required. The `refine` 73 | coroutine function is the opposite as the `prepare` coroutine function. It takes the data 74 | emitted by the model and reconstitutes it back into the same type of data that was originally 75 | received. This makes it easy to pipeline data in and out. 76 | 77 | Here is the `refine` function that goes along with this `prepare function`: 78 | 79 | 80 | .. code-block:: python 81 | 82 | async def refine(hub, pipe, data, preds): 83 | ''' 84 | Take the data from the model and refine it back into salt event format 85 | ''' 86 | dmap = hub.P[pipe] 87 | rets = [] 88 | for ind in range(len(preds)): 89 | if not preds[ind]: 90 | continue 91 | ret = {} 92 | for t_ind in range(len(dmap['tmap'])): 93 | ret[dmap['tmap'][t_ind]] = dmap['r_words'][data[ind][t_ind]] 94 | rets.append(ret) 95 | return rets 96 | 97 | The refine function takes the pipe, data and predictions. The predictions are the numbers that 98 | map to the dataset. So now we can just go over the dataset, line up the outliers and restore the 99 | data to what it was originally. 100 | -------------------------------------------------------------------------------- /docs/topics/egress.rst: -------------------------------------------------------------------------------- 1 | ====================== 2 | Writing Egress Plugins 3 | ====================== 4 | 5 | Egress plugins might be the easiest things to write. Just need to take the 6 | data refined by the data plugin and send it somewhere. The `egress` function will 7 | be called with each batch of data that gets sent out. The function is called `run` 8 | and needs to be a coroutine function. 9 | 10 | .. code-block:: python 11 | 12 | import pprint 13 | 14 | 15 | async def run(hub, pipe, data): 16 | ''' 17 | Print the egress data pipe to the cli 18 | ''' 19 | for comp in data: 20 | pprint.pprint(comp) 21 | 22 | This example is amazingly simple, but all you need to to is accept the pipe and the 23 | data and send that data somewhere. 24 | 25 | As always, accept the `hub` as the first argument so you have access to all of the 26 | data and plugins. 27 | -------------------------------------------------------------------------------- /docs/topics/flows.rst: -------------------------------------------------------------------------------- 1 | =================== 2 | Understanding Flows 3 | =================== 4 | 5 | The basic pattern used in AI/ML is Ingest -> Data Management -> AI/ML Model -> Output. 6 | Umbra seeks to make these stages pluggable and re-usable through a system we call `flows`. 7 | 8 | Flow Stages 9 | =========== 10 | 11 | The `Flow` defines how this pipeline can be executed. Taking input data from any pluggable 12 | source and then moving it through the process. Umbra breaks this process up into multiple 13 | stages. These stages are called `ingress`, `data`, `model`, `egress`. 14 | 15 | Ingress 16 | ------- 17 | 18 | The ingress system is used to attach to an ongoing ingress system. Typically an event 19 | based system. The event system will emit events as they occur and push them through the 20 | pipeline. The ingress system can be a single instance ingress system as well, say in the 21 | form of a json file. 22 | 23 | Data 24 | ---- 25 | 26 | The data stage is used to create the input and output data paths. For most AI systems all 27 | of the input data needs to be reduced to numbers. The data plugins are made to take arbitrary 28 | data types and reform them into standardized data sets. 29 | 30 | For instance the `salt_event` data plugin takes the information in the Salt event stream and 31 | converts the words into numbers dynamically. The word datasets are created on the fly allowing 32 | different matched event streams to be prepared. 33 | 34 | Model 35 | ----- 36 | 37 | The model is the meat of the process. This is the area where Umbra calls out to tools like 38 | Tensorflow, pyod, and Scikit. The model receives the conditioned data from the data stage 39 | and crunches it. The model will also determine if the data is to be used for training or 40 | for predictions based on options like `train_for`. 41 | 42 | Egress 43 | ------ 44 | 45 | Once the model has run it can emit predictions and suggestions. The Egress system allows for 46 | the suggestions to be emitted out on another event based system. This can be an alerting system, 47 | notifications, or just a datastore holding the information. 48 | 49 | Flows and Pipes 50 | =============== 51 | 52 | In the flow configurations you define pipes and each pipe has the options for the named stages, 53 | as well as additional options for the pipe. All of the pipes defined in the flow files need to 54 | have unique names. 55 | 56 | A flow file with a single pipe called 'sh' looks like this: 57 | 58 | .. code-block:: yaml 59 | 60 | sh: 61 | ingress: 62 | salt_event: 'salt/beacon/*/sh*' 63 | data: salt_event 64 | model: knn 65 | egress: salt_event 66 | train_for: 50000 67 | enabled: True 68 | 69 | This defines that we will be attaching to the salt event bus as our ingress point and looking 70 | for events that match the given tag. The data modifier to use is obviously `salt_event` because 71 | we are attached to the salt_event ingress system. In this case we are using the simple `knn` 72 | model for outlier detection. Finally the data will be emitted back on the `salt_event` system 73 | as well. 74 | 75 | The additional options here are `train_for` and `enabled`. The `train_for` option allows for 76 | setting a finite number of data entries to train on before running predictions. `enabled` 77 | allows you to enable or disable the given pipe. 78 | -------------------------------------------------------------------------------- /docs/topics/ingress.rst: -------------------------------------------------------------------------------- 1 | ======================= 2 | Writing Ingress Plugins 3 | ======================= 4 | 5 | Umbra uses the venerable Plugin Oriented Programing paradigm as it is realized 6 | in the `pop` framework. This means that all of the features of `pop` are 7 | available. 8 | 9 | Making an ingress plugin is easy, just add the plugin to `umbra/mods/ingress`. 10 | The `ingress` plugin subsystem only takes a single function, `run`. This can be 11 | a function of any type, a generator, a coroutine, an async generator or a 12 | standard function. The type of function you choose to implement defines how the 13 | ingress system will run. If the ingress system needs to be running continuously 14 | then an async generator is optimal. 15 | 16 | This is a simple example of using a subprocess to tail a file. This example shows 17 | how to use subprocess with asyncio in python to shell out to an event stream. There 18 | are better ways to do what we are doing here, but this is a good example of using 19 | subprocess to stream, as it is often a great way to ingest data. 20 | 21 | .. code-block:: python 22 | 23 | import asyncio 24 | import json 25 | 26 | async def run(hub): 27 | ''' 28 | Shell out salt-run and send in the events from the salt event bus 29 | ''' 30 | proc = await asyncio.create_subprocess_shell( 31 | 'salt-run state.event', 32 | stdout=asyncio.subprocess.PIPE) 33 | while True: 34 | line = await proc.stdout.readline() 35 | line = line.decode() 36 | comps = line.split(maxsplit=1) 37 | if len(comps) < 2: 38 | continue 39 | tag = comps[0].strip() 40 | data = json.loads(comps[1]) 41 | group = {} 42 | for pipe in conf: 43 | group[pipe] = [] 44 | for pipe in conf: 45 | for match in conf[pipe]: 46 | if fnmatch.fnmatch(tag, match): 47 | data = json.loads(comps[1].strip()) 48 | event = {'tag': tag, 'data': data} 49 | group[pipe].append(event) 50 | for pipe in group: 51 | yield {'pipe': pipe, 'data': group[pipe]} 52 | 53 | This example shows how to easily use asyncio subprocess to await lines by line 54 | feedback and yield the formatted ingestion data. This approach can work with 55 | virtually any shell program that continuously emits data. 56 | -------------------------------------------------------------------------------- /docs/topics/models.rst: -------------------------------------------------------------------------------- 1 | ====================== 2 | Writing Models Plugins 3 | ====================== 4 | 5 | The model is the heart of Umbra. This is where the AI/ML routines are executed with all 6 | of our carefully ingested and prepared datasets. The models allow for the datasets to 7 | be used to train and build up the AI used to make predictions. The challenge is, as 8 | always, to produce effective models. 9 | 10 | The model though, does not need to be complex, it only needs to be able to take in training 11 | and prediction data. 12 | 13 | As always, Umbra is made using `pop` to create the plugin architecture. So you must accept 14 | the `hub` as the first argument. The following arguments are `data` and `train`, these arguments 15 | are the datasets used to predict and train. Here is how the setup looks: 16 | 17 | .. code-block:: python 18 | 19 | 20 | from pyod.models.knn import KNN 21 | 22 | 23 | def __mod_init__(hub): 24 | hub.models.knn.COMPS = {} 25 | 26 | 27 | async def run(hub, pipe, data, train): 28 | ''' 29 | Run the knn algorith on the given dataset 30 | ''' 31 | if pipe not in hub.models.knn.COMPS: 32 | hub.models.knn.COMPS[pipe] = {'knn': KNN(contamination=0.01)} 33 | knn = hub.models.knn.COMPS[pipe]['knn'] 34 | if train: 35 | knn.fit(train) 36 | if data: 37 | knn.fit(data) 38 | return knn.predict(data) 39 | return [] 40 | 41 | This is a great example of some of the benefits of `pop`. When the module is first loaded 42 | we want to create a dict on the hub's namespace that we can use. The `__mod_init__` function 43 | is executed just once, when the module is first loaded. This allows us to set up data on 44 | the module's hub namespace. This is what you see when we set `hub.models.knn.COMPS`. This 45 | allows us to persist things like the `knn` object that we are training. 46 | 47 | As you can see, this is an extremely simple example! Think of the model plugin interface as 48 | just a doorway to hook into a larger model! 49 | -------------------------------------------------------------------------------- /docs/topics/quickstart.rst: -------------------------------------------------------------------------------- 1 | ================ 2 | Umbra Quickstart 3 | ================ 4 | 5 | 6 | Using Umbra to apply AI to a system is intended to be as easy as possible. When 7 | making an AI/ML system there are many considerations, not just the actual AI/ML 8 | systems that are going to be used. Umbra seeks to make these steps re-usable and 9 | easy to apply to multiple types of applications. 10 | 11 | Installation 12 | ============ 13 | 14 | To get started with Umbra, first install the application: 15 | 16 | .. code-block:: bash 17 | 18 | pip install penumbra 19 | 20 | Remember, umbra requires python 3.6 or later to run! 21 | 22 | Making a Flow 23 | ============= 24 | 25 | .. note:: 26 | 27 | Umbra can be run as any user on your system. This intro will assume 28 | that you are running as root for simplicity, but is running as a 29 | non-root user the configurations used here will be available 30 | in that user's home directory under `~/.umbra` 31 | 32 | The first thing to do is to make a flow. Umbra defines the AI pipeline 33 | in flow files. Start by making a flow file in `/etc/umbra/flows/shell.yml`: 34 | 35 | .. code-block:: yaml 36 | 37 | shell: 38 | ingress: 39 | json: [/root/shell.json] 40 | data: salt_event 41 | model: knn 42 | egress: cli 43 | train_for: 10000 44 | 45 | This will set up Umbra to read in the dataset from the file located at 46 | `/root/shell.json`. This is the ingress point, it will be sent through the `salt_event` 47 | data manager which makes the dataset readable for AI, then it will used the `knn` model 48 | to analize the data. Finally the egress system is the cli, so the results will just be 49 | spit out to the command line. `train_for` tells us how many data points to use for training 50 | before we start to predict if the following data points are anomalies. 51 | 52 | You have just defined an Umbra `Pipe` called `shell`. You can run multiple pipes at the same 53 | time and create as many flow files as you like. 54 | 55 | The `shell.json` file can be generated by running the `mkdata.py` script found in the 56 | demo directory. It will use your shell history to create a larger dataset. 57 | 58 | Now you can run Umbra and see the anomalies: 59 | 60 | .. code-block:: bash 61 | 62 | umbra 63 | 64 | Give it a few moments to train on the dataset, and then it will tell you what outliers 65 | it found in the generated `shell.json` data. -------------------------------------------------------------------------------- /docs/topics/salt.rst: -------------------------------------------------------------------------------- 1 | ===================== 2 | Integrating with Salt 3 | ===================== 4 | 5 | Umbra can integrate directly with a Salt Master to detect anomalies on the event bus. This 6 | can be very useful when combined with a Salt Reactor. Use the following `flow` with umbra to 7 | integrate with the `sh` beacon running on multiple minions: 8 | 9 | .. code-block:: yaml 10 | 11 | sh: 12 | ingress: 13 | salt_event: 'salt/beacon/*/sh*' 14 | data: salt_event 15 | model: knn 16 | egress: salt_event 17 | train_for: 10000 18 | 19 | This flow, saved to `/etc/umbra/flows/sh.yml` will now activate umbra to attach to the Salt 20 | Master event bus, gather the events that match the given tag, and then emit events for strange 21 | shell commands that are executed on the attached minions. 22 | 23 | Now run `umbra` on the Salt Master, as the same user the Salt Master is running as. Umbra 24 | will emit events onto the Salt Event Bus with the tag: `umbra/sh`. Now you can set up a 25 | reactor to handle the events as the come in! 26 | 27 | Please keep in mind that the `train_for` number might be too high, or too low based on your 28 | environment. 29 | 30 | Additional Beacons 31 | ================== 32 | 33 | The goal of Umbra and Salt is to make extra use of the beacon system in Salt. Take a few moments 34 | to review what beacons are currently in Salt, and consider what other beacons you could build that 35 | will integrate with Umbra. Remember that adding beacons to Salt is very easy and Umbra is built to 36 | automatically format and manage the data being sent back from beacons making it easy to 37 | apply AI/ML to virtually any situation. 38 | -------------------------------------------------------------------------------- /flow.yml: -------------------------------------------------------------------------------- 1 | shell: 2 | ingress: 3 | salt_event: 'salt/sh*' 4 | data: salt_event 5 | model: knn 6 | egress: salt_event 7 | 8 | net: 9 | ingress: 10 | salt_event: 'salt/sh*' 11 | data: salt_event 12 | model: lof 13 | egress: salt_event 14 | train_for: 10000 -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | pop 2 | pyod 3 | keras 4 | keras-applications 5 | keras-preprocessing 6 | tensorflow -------------------------------------------------------------------------------- /setup.py: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env python3 2 | # -*- coding: utf-8 -*- 3 | 4 | # Import python libs 5 | import os 6 | import sys 7 | import shutil 8 | 9 | from setuptools import setup, Command 10 | 11 | NAME = 'umbra' 12 | PNAME = 'penumbra' 13 | DESC = ('A system for making the use of AI easily plugged into interfaces') 14 | 15 | # Version info -- read without importing 16 | _locals = {} 17 | with open('{}/version.py'.format(NAME)) as fp: 18 | exec(fp.read(), None, _locals) 19 | VERSION = _locals['version'] 20 | SETUP_DIRNAME = os.path.dirname(__file__) 21 | if not SETUP_DIRNAME: 22 | SETUP_DIRNAME = os.getcwd() 23 | 24 | 25 | class Clean(Command): 26 | user_options = [] 27 | def initialize_options(self): 28 | pass 29 | 30 | def finalize_options(self): 31 | pass 32 | 33 | def run(self): 34 | for subdir in (NAME, 'tests'): 35 | for root, dirs, files in os.walk(os.path.join(os.path.dirname(__file__), subdir)): 36 | for dir_ in dirs: 37 | if dir_ == '__pycache__': 38 | shutil.rmtree(os.path.join(root, dir_)) 39 | 40 | 41 | def discover_packages(): 42 | modules = [] 43 | for package in (NAME, ): 44 | for root, _, files in os.walk(os.path.join(SETUP_DIRNAME, package)): 45 | pdir = os.path.relpath(root, SETUP_DIRNAME) 46 | modname = pdir.replace(os.sep, '.') 47 | modules.append(modname) 48 | return modules 49 | 50 | 51 | setup(name=PNAME, 52 | author='Thomas S Hatch', 53 | author_email='thatch@saltstack.com', 54 | url='https://saltstack.com', 55 | version=VERSION, 56 | description=DESC, 57 | classifiers=[ 58 | 'Operating System :: OS Independent', 59 | 'Programming Language :: Python', 60 | 'Programming Language :: Python :: 3.6', 61 | 'Programming Language :: Python :: 3.7', 62 | 'Development Status :: 5 - Production/Stable', 63 | ], 64 | entry_points={ 65 | 'console_scripts': [ 66 | 'umbra = umbra.scripts:start', 67 | ], 68 | }, 69 | packages=discover_packages(), 70 | cmdclass={'clean': Clean}, 71 | ) 72 | -------------------------------------------------------------------------------- /umbra/conf.py: -------------------------------------------------------------------------------- 1 | # Things to be aware of: 2 | # 1. The absolute paths get normalized when a non-root user is logged in, 3 | # so don't complain about it!' 4 | CLI_CONFIG = { 5 | 'flows_dir': { 6 | 'options': ['-f'], 7 | 'default': '/etc/umbra/flows', 8 | 'help': 'The location used for storing flow configuration files' 9 | }, 10 | 'cache_dir': { 11 | 'default': '/var/cache/umbra', 12 | 'help': 'The location for cache files', 13 | }, 14 | 'config': { 15 | 'options': ['-c'], 16 | 'default': '/etc/umbra/umbra.conf', 17 | 'help': 'The location to store umbra configuration files', 18 | }, 19 | 'persist': { 20 | 'options': ['-p'], 21 | 'default': '', 22 | 'help': 'Define the persistence system and options', 23 | }, 24 | } 25 | CONFIG = { 26 | 'flows_dir': { 27 | 'options': ['-f'], 28 | 'default': '/etc/umbra/flows', 29 | 'help': 'The location used for storing flow configuration files' 30 | }, 31 | 'cache_dir': { 32 | 'default': '/var/cache/umbra', 33 | 'help': 'The location for cache files', 34 | }, 35 | 'config': { 36 | 'options': ['-c'], 37 | 'default': '/etc/umbra/umbra.conf', 38 | 'help': 'The location to store umbra configuration files', 39 | }, 40 | 'persist': { 41 | 'options': ['-p'], 42 | 'default': '', 43 | 'help': 'Define the persistence system and options', 44 | }, 45 | } 46 | GLOBAL = {} 47 | SUBS = {} 48 | -------------------------------------------------------------------------------- /umbra/data/init.py: -------------------------------------------------------------------------------- 1 | async def run(hub, flows): 2 | ''' 3 | Take the flows that we want to run and start monitoring the respective data pipelines 4 | ''' 5 | for pipe in flows: 6 | hub.pop.loop.ensure_future('data.init.flow', pipe, flows[pipe]) 7 | 8 | 9 | async def flow(hub, pipe, config): 10 | ''' 11 | Execute the data processing for the given pipe 12 | ''' 13 | mod = config['data'] 14 | if pipe not in hub.P: 15 | hub.P[pipe] = { 16 | 'tmap': [], 17 | 'words': {}, 18 | 'r_words': {}, 19 | 'data': [], 20 | 'tmap_populated': False, 21 | 'first': False} 22 | # TODO: Make the flow stop if the ingress gets exhausted 23 | while True: 24 | data = await hub.UP[pipe]['data'].get() 25 | cond = await getattr(hub, f'data.{mod}.prepare')(pipe, data) 26 | await hub.UP[pipe]['model'].put(cond) 27 | 28 | 29 | def word_map(hub, word, words, r_words): 30 | ''' 31 | Take a word and build it into the dict map. Modify the words dicts in place and then 32 | return the int that represents the given word 33 | ''' 34 | if word in words: 35 | return words[word] 36 | num = len(words) + 1 37 | words[word] = num 38 | r_words[num] = word 39 | return num 40 | -------------------------------------------------------------------------------- /umbra/data/salt_event.py: -------------------------------------------------------------------------------- 1 | ''' 2 | Take the data from a salt event stream and prepare it to be loaded into an 3 | AI/ML model 4 | ''' 5 | 6 | 7 | async def prepare(hub, pipe, inbound): 8 | ''' 9 | Takes the raw data loaded from a salt event stream, this data is transformed 10 | into data that can be loaded in to the model and we save the string map dicts 11 | for future translation 12 | ''' 13 | ret = [] 14 | d_count = 0 15 | for event in inbound: 16 | if not hub.P[pipe]['tmap_populated']: 17 | for key in event['data']: 18 | if not isinstance(event['data'][key], str): 19 | # TODO: Make this able to handle more than just strings 20 | continue 21 | if key == '_stamp': 22 | continue 23 | hub.P[pipe]['tmap'].append(key) 24 | hub.P[pipe]['tmap_populated'] = True 25 | x = [0 for n in hub.P[pipe]['tmap']] 26 | for key in event['data']: 27 | if not isinstance(event['data'][key], str): 28 | # TODO: Be able to handle more than just strings here 29 | continue 30 | if key == '_stamp': 31 | continue 32 | x[hub.P[pipe]['tmap'].index(key)] = hub.data.init.word_map( 33 | event['data'][key], 34 | hub.P[pipe]['words'], 35 | hub.P[pipe]['r_words']) 36 | ret.append(x) 37 | return ret 38 | 39 | 40 | async def refine(hub, pipe, data, preds): 41 | ''' 42 | Take the data from the model and refine it back into salt event format 43 | ''' 44 | dmap = hub.P[pipe] 45 | rets = [] 46 | for ind in range(len(preds)): 47 | ret = {} 48 | for t_ind in range(len(dmap['tmap'])): 49 | ret[dmap['tmap'][t_ind]] = dmap['r_words'][data[ind][t_ind]] 50 | ret['pred'] = preds[ind] 51 | rets.append(ret) 52 | return rets 53 | -------------------------------------------------------------------------------- /umbra/egress/cli.py: -------------------------------------------------------------------------------- 1 | # Import python libs 2 | import pprint 3 | 4 | 5 | async def run(hub, pipe, data): 6 | ''' 7 | Print the egress data pipe to the cli 8 | ''' 9 | for comp in data: 10 | pprint.pprint(comp) 11 | print(len(data)) -------------------------------------------------------------------------------- /umbra/egress/init.py: -------------------------------------------------------------------------------- 1 | async def run(hub, flows): 2 | ''' 3 | Take the egress data flow, re-normalize the data and then push predictions out 4 | ''' 5 | for pipe in flows: 6 | hub.pop.loop.ensure_future('egress.init.flow', pipe, flows[pipe]) 7 | 8 | 9 | async def flow(hub, pipe, conf): 10 | ''' 11 | Take the given pipe and flow and execute it 12 | ''' 13 | e_mod = conf['egress'] 14 | d_mod = conf['data'] 15 | if not isinstance(e_mod, list): 16 | e_mod = [e_mod] 17 | while True: 18 | w_preds = await hub.UP[pipe]['egress'].get() 19 | data = await getattr(hub, f'data.{d_mod}.refine')( 20 | pipe, 21 | w_preds['data'], 22 | w_preds['preds']) 23 | for mod in e_mod: 24 | await getattr(hub, f'egress.{mod}.run')(pipe, data) 25 | -------------------------------------------------------------------------------- /umbra/egress/salt_event.py: -------------------------------------------------------------------------------- 1 | # import python libs 2 | import asyncio 3 | import json 4 | 5 | 6 | async def run(hub, pipe, data): 7 | ''' 8 | Emit events onto the salt event bus that signal umbra predictions 9 | ''' 10 | tag = f'umbra/{pipe}' 11 | for comp in data: 12 | proc = await asyncio.create_subprocess_shell( 13 | f'salt-run event.send {tag} \'{json.dumps(comp)}\'', 14 | stdout=asyncio.subprocess.PIPE, 15 | stderr=asyncio.subprocess.PIPE) 16 | await proc.wait() -------------------------------------------------------------------------------- /umbra/flows/init.py: -------------------------------------------------------------------------------- 1 | # Import python libs 2 | import os 3 | # Import third party libs 4 | import yaml 5 | 6 | ''' 7 | # Flows file format 8 | shell: 9 | ingress: 10 | salt_event: 'salt/sh*' 11 | data: salt_event 12 | model: knn 13 | egress: salt_event 14 | 15 | net: 16 | ingress: 17 | salt_event: 'salt/sh*' 18 | data: salt_event 19 | model: k_means 20 | egress: salt_event 21 | ''' 22 | 23 | 24 | def load(hub): 25 | ''' 26 | Read in the files that reside in the flows direrctory and merge them into 27 | the flow map used to run the entire process 28 | ''' 29 | flows = {} 30 | f_dir = hub.OPT['umbra']['flows_dir'] 31 | for fn in os.listdir(f_dir): 32 | full = os.path.join(f_dir, fn) 33 | with open(full, 'r') as rfh: 34 | data = yaml.safe_load(rfh.read()) 35 | flows[full] = data 36 | #TODO: FIx this namespace to be hub.flows 37 | hub.umbra.INGRESS, hub.umbra.FLOWS = hub.flows.init.merge(flows) 38 | 39 | 40 | def merge(hub, flows): 41 | ''' 42 | Given the flows, find the areas where they converge and merge the related 43 | ingress data 44 | ''' 45 | ingress = {} 46 | r_flows = {} 47 | for full, data in flows.items(): 48 | for pipe in data: 49 | if not data[pipe].get('enabled', True): 50 | continue 51 | for key in data[pipe]['ingress']: 52 | if key not in ingress: 53 | ingress[key] = {} 54 | if pipe not in ingress[key]: 55 | ingress[key][pipe] = [] 56 | val = data[pipe]['ingress'][key] 57 | if not isinstance(val, list): 58 | ingress[key][pipe].append(val) 59 | else: 60 | ingress[key][pipe].extend(val) 61 | r_flows[pipe] = data[pipe] 62 | return ingress, r_flows 63 | -------------------------------------------------------------------------------- /umbra/ingress/init.py: -------------------------------------------------------------------------------- 1 | # Import python libs 2 | import asyncio 3 | import types 4 | 5 | 6 | async def run(hub, ingress): 7 | ''' 8 | Start up the coroutines for each ingress pipe 9 | ''' 10 | for mod, conf in ingress.items(): 11 | for pipe in conf: 12 | if pipe not in hub.UP: 13 | hub.UP[pipe] = {} 14 | hub.UP[pipe]['in'] = asyncio.Queue() 15 | hub.UP[pipe]['data'] = asyncio.Queue() 16 | hub.UP[pipe]['model'] = asyncio.Queue() 17 | hub.UP[pipe]['egress'] = asyncio.Queue() 18 | hub.pop.loop.ensure_future('ingress.init.flow', mod, conf) 19 | 20 | 21 | async def flow(hub, mod, conf): 22 | ret = getattr(hub, f'ingress.{mod}.run')(conf) 23 | if asyncio.iscoroutine(ret): 24 | ret = await ret 25 | await hub.ingress.init.que(ret) 26 | elif isinstance(ret, types.AsyncGeneratorType): 27 | async for chunk in ret: 28 | await hub.ingress.init.que(chunk) 29 | elif isinstance(ret, types.GeneratorType): 30 | for chunk in ret: 31 | await hub.ingress.init.que(chunk) 32 | else: 33 | await hub.ingress.init.que(ret) 34 | 35 | 36 | async def que(hub, chunk): 37 | ''' 38 | Take a single return from the ingress functions 39 | ''' 40 | pipe = chunk['pipe'] 41 | data = chunk['data'] 42 | await hub.UP[pipe]['data'].put(data) 43 | -------------------------------------------------------------------------------- /umbra/ingress/json_file.py: -------------------------------------------------------------------------------- 1 | # Import python libs 2 | import json 3 | 4 | __virtualname__ = 'json' 5 | 6 | 7 | async def run(hub, conf): 8 | ''' 9 | Read in the configured json files full of juicy data. Add those files' data to the named pipe 10 | ''' 11 | for pipe in conf: 12 | for fn in conf[pipe]: 13 | with open(fn, 'r') as rfh: 14 | data = json.loads(rfh.read()) 15 | yield {'pipe': pipe, 'data': data} -------------------------------------------------------------------------------- /umbra/ingress/salt_event.py: -------------------------------------------------------------------------------- 1 | # Import python libs 2 | import asyncio 3 | import fnmatch 4 | import json 5 | 6 | 7 | async def run(hub, conf): 8 | ''' 9 | Take the available events off the bus, match them, and then place them on the correct pipes 10 | ''' 11 | proc = await asyncio.create_subprocess_shell( 12 | 'salt-run state.event', 13 | stdout=asyncio.subprocess.PIPE) 14 | while True: 15 | line = await proc.stdout.readline() 16 | line = line.decode() 17 | comps = line.split(maxsplit=1) 18 | if len(comps) < 2: 19 | continue 20 | tag = comps[0].strip() 21 | data = json.loads(comps[1]) 22 | group = {} 23 | for pipe in conf: 24 | group[pipe] = [] 25 | for pipe in conf: 26 | for match in conf[pipe]: 27 | if fnmatch.fnmatch(tag, match): 28 | data = json.loads(comps[1].strip()) 29 | event = {'tag': tag, 'data': data} 30 | group[pipe].append(event) 31 | for pipe in group: 32 | yield {'pipe': pipe, 'data': group[pipe]} -------------------------------------------------------------------------------- /umbra/models/auto_encoder.py: -------------------------------------------------------------------------------- 1 | ''' 2 | Take dataset X and run it through the knn algorithm 3 | ''' 4 | 5 | # Import third party libs 6 | from pyod.models.auto_encoder import AutoEncoder 7 | 8 | __virtualname__ = 'auto_encoder' 9 | 10 | 11 | def __init__(hub): 12 | hub.models.auto_encoder.COMPS = {} 13 | 14 | 15 | def make_mlo(hub, data, train): 16 | ''' 17 | Create the Machine Learning Object used for this sequence 18 | ''' 19 | size = 0 20 | for chunk in data: 21 | size = len(chunk) 22 | break 23 | for chunk in train: 24 | size = len(chunk) 25 | break 26 | hidden_neurons = [size*2, size, size, size*2] 27 | return AutoEncoder(hidden_neurons=hidden_neurons, contamination=0.001) 28 | 29 | 30 | async def run(hub, config, pipe, data, train): 31 | ''' 32 | Run the auto_encoder algorithm on the given dataset 33 | ''' 34 | if pipe not in hub.models.auto_encoder.COMPS: 35 | hub.models.auto_encoder.COMPS[pipe] = {'mlo': hub.models.auto_encoder.make_mlo(data, train)} 36 | mlo = hub.models.auto_encoder.COMPS[pipe]['mlo'] 37 | if train: 38 | mlo.fit(train) 39 | if data: 40 | mlo.fit(data) 41 | return mlo.predict(data) 42 | return [] 43 | -------------------------------------------------------------------------------- /umbra/models/hbos.py: -------------------------------------------------------------------------------- 1 | ''' 2 | Take dataset X and run it through the knn algorithm 3 | ''' 4 | 5 | # Import third party libs 6 | from pyod.models.hbos import HBOS 7 | 8 | __virtualname__ = 'hbos' 9 | 10 | 11 | def __init__(hub): 12 | hub.models.hbos.COMPS = {} 13 | 14 | 15 | def make_mlo(hub, data, train): 16 | ''' 17 | Create the Machine Learning Object used for this sequence 18 | ''' 19 | return HBOS(contamination=0.001) 20 | 21 | 22 | async def run(hub, config, pipe, data, train): 23 | ''' 24 | Run the hbos algorithm on the given dataset 25 | ''' 26 | if pipe not in hub.models.hbos.COMPS: 27 | hub.models.hbos.COMPS[pipe] = {'mlo': hub.models.hbos.make_mlo(data, train)} 28 | mlo = hub.models.hbos.COMPS[pipe]['mlo'] 29 | if train: 30 | mlo.fit(train) 31 | if data: 32 | mlo.fit(data) 33 | return mlo.predict(data) 34 | return [] 35 | -------------------------------------------------------------------------------- /umbra/models/init.py: -------------------------------------------------------------------------------- 1 | async def run(hub, flows): 2 | ''' 3 | Execute the models defined in the given flow 4 | ''' 5 | hub.models.TRAIN = {} 6 | for pipe in flows: 7 | hub.models.TRAIN[pipe] = 0 8 | hub.pop.loop.ensure_future('models.init.flow', pipe, flows[pipe]) 9 | 10 | 11 | async def flow(hub, pipe, config): 12 | ''' 13 | Given the config, fire up the model pulling in the data from the respective 14 | data pipe 15 | ''' 16 | mod = config['model'] 17 | print(f'Starting model: {mod}') 18 | data = [] 19 | train = [] 20 | while True: 21 | data.extend(await hub.UP[pipe]['model'].get()) 22 | if hub.P[pipe]['first']: 23 | train.extend(hub.P[pipe]['data']) 24 | hub.P[pipe]['first'] = False 25 | if config.get('train_for'): 26 | tnum = config['train_for'] 27 | left = tnum - hub.models.TRAIN[pipe] - len(train) 28 | if left > 0: 29 | # Move the numberd items from data into train 30 | # TODO: We can do this better using more math and slices 31 | while left > 0: 32 | if data: 33 | train.append(data.pop(0)) 34 | left -= 1 35 | else: 36 | break 37 | if data and len(data) < config.get('group_size', 10): 38 | continue 39 | if train and len(train) < config.get('train_for'): 40 | if not len(train) % 100: 41 | print(f'Training data prepared: {len(train)}') 42 | continue 43 | hub.models.TRAIN[pipe] += len(train) 44 | if hub.OPT['umbra']['persist']: 45 | # TODO: This is a memory leak. We need to store this seperately and not keep it all in ram 46 | hub.P[pipe]['data'].extend(data) 47 | preds = await getattr(hub, f'models.{mod}.run')(config, pipe, data, train) 48 | if hub.OPT['umbra']['persist']: 49 | await hub.persist.init.dump() 50 | await hub.UP[pipe]['egress'].put({'data': data, 'preds': preds}) 51 | data = [] 52 | train = [] 53 | -------------------------------------------------------------------------------- /umbra/models/kmeans.py: -------------------------------------------------------------------------------- 1 | ''' 2 | Take dataset X and run it through the k-means algorithm 3 | 4 | Flow config: 5 | 6 | model: kmeans 7 | kmeans: 8 | n_clusters: 20 9 | n_init: 10 10 | max_iter: 500 11 | ''' 12 | 13 | # Import third party libs 14 | from sklearn.cluster import KMeans 15 | 16 | 17 | def __init__(hub): 18 | hub.models.kmeans.COMPS = {} 19 | 20 | 21 | async def run(hub, config, pipe, data, train): 22 | ''' 23 | Run the k-means algorithm on the given dataset 24 | ''' 25 | if pipe not in hub.models.kmeans.COMPS: 26 | kmconfig = config.get('kmeans', {}) 27 | mlo = KMeans(n_clusters=kmconfig.get('n_clusters', 8), 28 | n_init=kmconfig.get('n_init', 10), 29 | max_iter=kmconfig.get('max_iter', 300)) 30 | print('Created k-means machine learning object:\n', mlo) 31 | hub.models.kmeans.COMPS[pipe] = {'mlo': mlo} 32 | 33 | mlo = hub.models.kmeans.COMPS[pipe]['mlo'] 34 | if train: 35 | print(f'Training {len(train)} samples') 36 | mlo.fit(train) 37 | if data: 38 | print(f'Predicting {len(data)} samples') 39 | return list(mlo.predict(data)) 40 | return [] 41 | -------------------------------------------------------------------------------- /umbra/models/knn.py: -------------------------------------------------------------------------------- 1 | ''' 2 | Take dataset X and run it through the knn algorithm 3 | ''' 4 | 5 | # Import third party libs 6 | from pyod.models.knn import KNN 7 | 8 | 9 | def __init__(hub): 10 | hub.models.knn.COMPS = {} 11 | 12 | 13 | async def run(hub, config, pipe, data, train): 14 | ''' 15 | Run the knn algorith on the given dataset 16 | ''' 17 | if pipe not in hub.models.knn.COMPS: 18 | hub.models.knn.COMPS[pipe] = {'clf': KNN(contamination=0.01)} 19 | clf = hub.models.knn.COMPS[pipe]['clf'] 20 | if train: 21 | print(f'Training {len(train)} datasets') 22 | clf.fit(train) 23 | if data: 24 | print(f'Fitting {len(data)} datasets') 25 | clf.fit(data) 26 | print(f'Predicting {len(data)} datasets') 27 | return clf.predict(data) 28 | return [] 29 | -------------------------------------------------------------------------------- /umbra/models/lof.py: -------------------------------------------------------------------------------- 1 | ''' 2 | Take dataset X and run it through the lof algorithm 3 | ''' 4 | 5 | # Import third party libs 6 | from pyod.models.lof import LOF 7 | 8 | __virtualname__ = 'lof' 9 | 10 | 11 | def __init__(hub): 12 | hub.models.lof.COMPS = {} 13 | 14 | 15 | def make_mlo(hub, data, train): 16 | ''' 17 | Create the Machine Learning Object used for this sequence 18 | ''' 19 | return LOF(contamination=0.01) 20 | 21 | 22 | async def run(hub, config, pipe, data, train): 23 | ''' 24 | Run the lof algorithm on the given dataset 25 | ''' 26 | if pipe not in hub.models.lof.COMPS: 27 | hub.models.lof.COMPS[pipe] = {'mlo': hub.models.lof.make_mlo(data, train)} 28 | mlo = hub.models.lof.COMPS[pipe]['mlo'] 29 | if train: 30 | print(f'Training {len(train)} datasets') 31 | mlo.fit(train) 32 | if data: 33 | print(f'Predicting {len(data)} datasets') 34 | ret = mlo.predict(data) 35 | scores = mlo.decision_function(data) 36 | return ret 37 | return [] 38 | -------------------------------------------------------------------------------- /umbra/models/lscp.py: -------------------------------------------------------------------------------- 1 | ''' 2 | Take dataset X and run it through the knn algorithm 3 | ''' 4 | 5 | # Import third party libs 6 | from pyod.models.lscp import LSCP 7 | 8 | __virtualname__ = 'lscp' 9 | 10 | 11 | def __init__(hub): 12 | hub.models.lscp.COMPS = {} 13 | 14 | 15 | def make_mlo(hub, data, train): 16 | ''' 17 | Create the Machine Learning Object used for this sequence 18 | ''' 19 | return LSCP(contamination=0.001) 20 | 21 | 22 | async def run(hub, config, pipe, data, train): 23 | ''' 24 | Run the lscp algorithm on the given dataset 25 | ''' 26 | if pipe not in hub.models.lscp.COMPS: 27 | hub.models.lscp.COMPS[pipe] = {'mlo': hub.models.lscp.make_mlo(data, train)} 28 | mlo = hub.models.lscp.COMPS[pipe]['mlo'] 29 | if train: 30 | mlo.fit(train) 31 | if data: 32 | mlo.fit(data) 33 | return mlo.predict(data) 34 | return [] 35 | -------------------------------------------------------------------------------- /umbra/models/optics.py: -------------------------------------------------------------------------------- 1 | ''' 2 | Take dataset X and run it through the OPTICS algorithm 3 | 4 | Flow config: 5 | 6 | model: optics 7 | optics: 8 | n_jobs: -1 9 | ''' 10 | 11 | # Import third party libs 12 | from sklearn.cluster import OPTICS 13 | 14 | 15 | def __init__(hub): 16 | hub.models.optics.COMPS = {} 17 | 18 | 19 | async def run(hub, config, pipe, data, train): 20 | ''' 21 | Run the OPTICS algorithm on the given dataset 22 | ''' 23 | if pipe not in hub.models.optics.COMPS: 24 | kmconfig = config.get('optics', {}) 25 | mlo = OPTICS(n_jobs=kmconfig.get('n_jobs', -1)) 26 | print('Created OPTICS machine learning object:\n', mlo) 27 | hub.models.optics.COMPS[pipe] = {'mlo': mlo} 28 | 29 | mlo = hub.models.optics.COMPS[pipe]['mlo'] 30 | if train: 31 | print(f'Training {len(train)} samples') 32 | mlo.fit(train) 33 | if data: 34 | print(f'Predicting {len(data)} samples') 35 | return list(mlo.fit_predict(data)) 36 | return [] 37 | -------------------------------------------------------------------------------- /umbra/persist/init.py: -------------------------------------------------------------------------------- 1 | async def load(hub): 2 | ''' 3 | Look to the configured persistence system and load up the most recent 4 | dataset into the persist system 5 | ''' 6 | p_name = hub.OPT['umbra']['persist'] 7 | hub.P = await hub.pop.ref.last(f'persist.{p_name}.load')() 8 | # We need to set the pipe's flag to let the ingestion know to re-train on historic data 9 | for pipe in hub.P: 10 | hub.P[pipe]['first'] = True 11 | 12 | 13 | async def dump(hub): 14 | ''' 15 | Take the current data from the runnign pipes and save it 16 | ''' 17 | p_name = hub.OPT['umbra']['persist'] 18 | await hub.pop.ref.last(f'persist.{p_name}.dump')(hub.P) 19 | -------------------------------------------------------------------------------- /umbra/persist/msgpack_file.py: -------------------------------------------------------------------------------- 1 | # Import python libs 2 | import os 3 | 4 | # Import third party libs 5 | import msgpack 6 | 7 | __virtualname__ = 'msgpack' 8 | 9 | 10 | async def load(hub): 11 | ''' 12 | Load up the data from the msgpack file 13 | ''' 14 | path = os.path.join(hub.OPT['umbra']['cache_dir'], 'data.mp') 15 | if os.path.isfile(path): 16 | with open(path, 'rb') as rfh: 17 | return msgpack.loads(rfh.read()) 18 | else: 19 | return {} 20 | 21 | 22 | async def dump(hub, data): 23 | ''' 24 | Dump the persistence running data to disk 25 | ''' 26 | path = os.path.join(hub.OPT['umbra']['cache_dir'], 'data.mp') 27 | with open(path, 'wb+') as wfh: 28 | wfh.write(msgpack.dumps(data)) 29 | 30 | -------------------------------------------------------------------------------- /umbra/scripts.py: -------------------------------------------------------------------------------- 1 | # Import pop libs 2 | import pop.hub 3 | 4 | def start(): 5 | hub = pop.hub.Hub() 6 | hub.pop.sub.add(pypath='umbra.umbra') 7 | -------------------------------------------------------------------------------- /umbra/umbra/init.py: -------------------------------------------------------------------------------- 1 | # Import python libs 2 | import asyncio 3 | 4 | # Application data flow: 5 | # Flows saved to hub.umbra.INGRESS and hub.umbra.FLOWS 6 | # Data pipe saved to hub.UP 7 | # hub.UP[pipe][in] 8 | # hub.UP[pipe][data] 9 | # hub.UP[pipe][model] 10 | # hub.UP[pipe][persist] 11 | # hub.UP[pipe][egress] 12 | 13 | 14 | def __init__(hub): 15 | hub.pop.conf.integrate(['umbra'], loader='yaml', cli='umbra', roots=True) 16 | hub.UP = {} 17 | hub.P = {} 18 | hub.umbra.init.load_subs() 19 | hub.flows.init.load() 20 | hub.umbra.init.start() 21 | 22 | 23 | def load_subs(hub): 24 | hub.pop.sub.add(pypath='umbra.flows') 25 | hub.pop.sub.add(pypath='umbra.persist') 26 | hub.pop.sub.add(pypath='umbra.ingress') 27 | hub.pop.sub.add(pypath='umbra.data') 28 | hub.pop.sub.add(pypath='umbra.models') 29 | hub.pop.sub.add(pypath='umbra.egress') 30 | 31 | 32 | def start(hub): 33 | ''' 34 | Fire up the async loop and add the first coroutine 35 | ''' 36 | hub.pop.loop.start( 37 | hub.umbra.init.run(), 38 | hold=True) 39 | 40 | 41 | async def run(hub): 42 | ''' 43 | Start up the flows process 44 | ''' 45 | if hub.OPT['umbra']['persist']: 46 | await hub.persist.init.load() 47 | hub.umbra.INBOUND = asyncio.Queue() 48 | await hub.ingress.init.run(hub.umbra.INGRESS) 49 | await hub.data.init.run(hub.umbra.FLOWS) 50 | await hub.models.init.run(hub.umbra.FLOWS) 51 | await hub.egress.init.run(hub.umbra.FLOWS) 52 | -------------------------------------------------------------------------------- /umbra/version.py: -------------------------------------------------------------------------------- 1 | # All pop projects follow semantic versioning version 2.0.0: https://semver.org/ 2 | version = '1.3.0' 3 | --------------------------------------------------------------------------------