2 | This allows you to specify a mathematical expression which, if true, will turn the job to unstbale
3 | First the expression is split by logical operands of |,or,&,and,impl,imp,eq
4 | Then it it is split by >,<, ==, !=, <=, >=, lt, gt, le, ge
5 | Negation can be used only with logicalbrackets, strictly tied to left one - eg ![0 == 1]
6 | each of the remaining parts is calcualted as mathematical expression. as powerfull as https://github.com/gbenroscience/ParserNG, with few extensions:
7 | you can use L0 to access value of just finished build. L1 as value of last one before, L2 as two before and so on...Negative id on L is not allowed nowhere
8 | points in chart chart with 4 bullets, wehich fifth build jsut finished would be L4 L3 L2 L1 L0
9 | you can use ranges by `..`. Eg ..L1 is all before L0 (thus Ln...L2,L1). Similarly, L3.. would be L3,L2,L1. You can use also both limits - eg: L1..L3 o
10 | Note, that the evaluation counts with what it see - if you denylist/allowlist, the points are filtered before given to formula. Similarly, if you show just last two point, you will get access only to L0 and L1. So show more :)
11 |
12 | Expression : [[ avg(..L1)*1.1 < L0 ] || [L1*1.3 < L0 ]] || [ avgN(count(..L0)/4, ..L1)*1.1
13 | Upon : 60,20,45,70
14 | As : Ln...L1,L0
15 | Expanded as: [[ avg(60,20,45)*1.1 < 70 ] || [45*1.3 < 70 ]] || [ avgN(count(60,20,45,70)/4, 60,20,45)*1.1<70 ]
16 |
17 | brackets: [[ avg(60,20,45)*1.1 < 70 ] || [45*1.3 < 70 ]] || [ avgN(count(60,20,45,70)/4, 60,20,45)*1.1<70 ]
18 |
19 | brackets: [ avg(60,20,45)*1.1 < 70 ] || [45*1.3 < 70 ]
20 |
21 | evaluating: avg(60,20,45)*1.1 < 70
22 |
23 | evaluating: avg(60,20,45)*1.1 < 70
24 |
25 | evaluating: avg(60,20,45)*1.1
26 | is: 45.833333333333336
27 | evaluating: 70
28 | is: 70
29 |
30 | ... 45.833333333333336 < 70
31 | is: true
32 |
33 | is: true
34 |
35 | to: true || [45*1.3 < 70 ]
36 |
37 | evaluating: 45*1.3 < 70
38 |
39 | evaluating: 45*1.3 < 70
40 |
41 | evaluating: 45*1.3
42 | is: 58.5
43 | evaluating: 70
44 | is: 70
45 |
46 | ... 58.5 < 70
47 | is: true
48 |
49 | is: true
50 |
51 | to: true || true
52 |
53 | evaluating: true || true
54 |
55 | evaluating: true
56 | is: true
57 | evaluating: true
58 | is: true
59 |
60 | ... true | true
61 | is: true
62 |
63 | true
64 |
65 | to: true || [ avgN(count(60,20,45,70)/4, 60,20,45)*1.1<70 ]
66 |
67 | evaluating: avgN(count(60,20,45,70)/4, 60,20,45)*1.1<70
68 |
69 | evaluating: avgN(count(60,20,45,70)/4, 60,20,45)*1.1<70
70 |
71 | evaluating: avgN(count(60,20,45,70)/4, 60,20,45)*1.1
72 | is: 45.833333337
73 | evaluating: 70
74 | is: 70
75 |
76 | ... 45.833333337 < 70
77 | is: true
78 |
79 | is: true
80 |
81 | to: true || true
82 |
83 | evaluating: true || true
84 |
85 | evaluating: true
86 | is: true
87 | evaluating: true
88 | is: true
89 |
90 | ... true | true
91 | is: true
92 |
93 | true
94 |
95 | is: true
96 |
97 | TRUE , thus job WILL be turned to unstable
98 | Max number in L (and thus also in functions) is 99
99 | The parser is far from being perfect, be nice to it.
100 | The () brackets are avaiable only in side math expressions. The Logical part is using [] as separators. Eg:
101 |
102 | 1+1 < (2+0)*1 impl [ [5 == 6 || 33<(22-20)*2 ]xor [ [ 5-3 < 2 or 7*(5+2)<=5 ] and 1+1 == 2]] eq [ true && false ]
103 |
104 |
105 | There is a special element MN, which represents count of input points, so you do not need to call `count(..L0)` arround and arround. So...
106 | Expression : avg(..L1)*1.1-MN < L0 | L1*1.3 + MN< L0
107 | Upon : 60,20,80,70
108 | As : Ln...L1,L0
109 | MN = 4
110 | Expanded as: avg(60,20,80)*1.1-4 < 70 | 80*1.3 + 4< 70
111 | ...indeed.
112 |
113 |
114 | Dynamic indexes
115 | You can calculate the Lsomething by expressions. To do that, use L{expression}. Eg:
116 |
117 | avg( ..L{MN/2}) < avg(L{MN/2}..)
118 | Expression : avg( ..L{MN/2}) < avg(L{MN/2}..)
119 | Upon : 2,4,6
120 | As : Ln...L1,L0
121 | MN = 3
122 | L indexes brackets: avg( ..L{3/2}) < avg(L{3/2}..)
123 | Expression : 3/2
124 | Expanded as: 3/2
125 | is: 1.5
126 | 3/2 = 1 (1.5)
127 | to: avg( ..L 1 ) < avg(L{3/2}..)
128 | Expression : 3/2
129 | Expanded as: 3/2
130 | is: 1.5
131 | 3/2 = 1 (1.5)
132 | to: avg( ..L 1 ) < avg(L 1 ..)
133 | Expanded as: avg( 2,4) < avg(4,6)
134 | avg( 2,4) < avg(4,6)
135 | brackets: avg( 2,4) < avg(4,6)
136 | evaluating: avg( 2,4) < avg(4,6)
137 | evaluating: avg( 2,4) < avg(4,6)
138 | evaluating: avg( 2,4)
139 | is: 3.0
140 | evaluating: avg(4,6)
141 | is: 5.0
142 | ... 3.0 < 5.0
143 | is: true
144 | is: true
145 | true
146 | is: true
147 | true
148 |
149 |
150 |
151 |
Preset equations
152 | There are some preset equations, usable in form of ID arg1 arg2 ... arg9. eg:
153 |
154 | FINAL_DOWN_CUTTING_OK 2 5 5 5
155 |
156 | To print them, use LIST_INTERNALS instead of expression or look onto
157 |
158 | pre pprepared availabe embedded functions and exempalr config
159 | You can add yours own preset expressions in main settings as url to file with definitions or as definitions themselves.
160 | The syntax is as you saw
161 |
162 | above:
163 |
164 | - empty lines ignored
165 | - behind hash(#) are comments and examples
166 | - First comment si mandatory, and is ID
167 | - after comments follow expression itself as by
168 | ParserNG rules
169 |
170 | - It can be multilined for readability (as ParserNG's --trim)
171 | - parameters are included substitued into expression via /*1*/ for first param, /*2*/ for second and os on
172 |
173 |
174 |