llumo 0.2.28__py3-none-any.whl → 0.2.29__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- llumo/google.py +16 -16
- llumo/openai.py +137 -130
- {llumo-0.2.28.dist-info → llumo-0.2.29.dist-info}/METADATA +1 -1
- {llumo-0.2.28.dist-info → llumo-0.2.29.dist-info}/RECORD +7 -7
- {llumo-0.2.28.dist-info → llumo-0.2.29.dist-info}/WHEEL +0 -0
- {llumo-0.2.28.dist-info → llumo-0.2.29.dist-info}/licenses/LICENSE +0 -0
- {llumo-0.2.28.dist-info → llumo-0.2.29.dist-info}/top_level.txt +0 -0
llumo/google.py
CHANGED
|
@@ -15,9 +15,9 @@ def evaluate_multiple(data, api_key=None, evals=["Response Correctness"]):
|
|
|
15
15
|
|
|
16
16
|
|
|
17
17
|
class ChatCompletionWithEval:
|
|
18
|
-
def __init__(self, response, evaluation):
|
|
18
|
+
def __init__(self, response, evaluation=None):
|
|
19
19
|
self._response = response
|
|
20
|
-
self.evaluation = evaluation
|
|
20
|
+
# self.evaluation = evaluation
|
|
21
21
|
|
|
22
22
|
def __getattr__(self, name):
|
|
23
23
|
return getattr(self._response, name)
|
|
@@ -46,21 +46,21 @@ class genai:
|
|
|
46
46
|
response = self._model_instance.generate_content(contents=contents, **kwargs)
|
|
47
47
|
output = response.text
|
|
48
48
|
|
|
49
|
-
eval_input = [{
|
|
50
|
-
|
|
51
|
-
|
|
52
|
-
|
|
53
|
-
}]
|
|
49
|
+
# eval_input = [{
|
|
50
|
+
# "query": contents,
|
|
51
|
+
# "context": context or contents,
|
|
52
|
+
# "output": output,
|
|
53
|
+
# }]
|
|
54
54
|
|
|
55
|
-
evaluation = None
|
|
56
|
-
try:
|
|
57
|
-
|
|
58
|
-
except Exception as e:
|
|
59
|
-
|
|
55
|
+
# evaluation = None
|
|
56
|
+
# try:
|
|
57
|
+
# evaluation = evaluate_multiple(data=eval_input, evals=evals, api_key=llumo_key)
|
|
58
|
+
# except Exception as e:
|
|
59
|
+
# evaluation = None
|
|
60
60
|
|
|
61
|
-
if evaluation is None:
|
|
62
|
-
|
|
63
|
-
|
|
61
|
+
# if evaluation is None:
|
|
62
|
+
# print("Cannot process your request for evaluation, please check your api and try again later.")
|
|
63
|
+
# return response
|
|
64
64
|
|
|
65
65
|
|
|
66
|
-
return ChatCompletionWithEval(response, evaluation)
|
|
66
|
+
return ChatCompletionWithEval(response, evaluation=None)
|
llumo/openai.py
CHANGED
|
@@ -25,9 +25,13 @@ def performEvaluation(data, api_key=None, evals=["Response Correctness"], **kwar
|
|
|
25
25
|
|
|
26
26
|
# Wrapper around ChatCompletion to allow custom fields like `.evaluation`
|
|
27
27
|
class ChatCompletionWithEval:
|
|
28
|
-
def __init__(
|
|
28
|
+
def __init__(
|
|
29
|
+
self,
|
|
30
|
+
response,
|
|
31
|
+
# , evaluation
|
|
32
|
+
):
|
|
29
33
|
self._response = response
|
|
30
|
-
self.evaluation = evaluation
|
|
34
|
+
# self.evaluation = evaluation
|
|
31
35
|
|
|
32
36
|
def __getattr__(self, name):
|
|
33
37
|
return getattr(self._response, name)
|
|
@@ -70,65 +74,65 @@ class OpenAI(OpenAIClient):
|
|
|
70
74
|
workspace_id = self.session.logger.getWorkspaceID()
|
|
71
75
|
|
|
72
76
|
# Input Bias Evaluation
|
|
73
|
-
eval_input_bias = [
|
|
74
|
-
|
|
75
|
-
|
|
76
|
-
|
|
77
|
-
|
|
78
|
-
|
|
79
|
-
]
|
|
80
|
-
try:
|
|
81
|
-
|
|
82
|
-
|
|
83
|
-
|
|
84
|
-
|
|
85
|
-
|
|
86
|
-
|
|
87
|
-
|
|
88
|
-
|
|
89
|
-
|
|
90
|
-
|
|
91
|
-
|
|
92
|
-
|
|
93
|
-
|
|
94
|
-
|
|
95
|
-
|
|
96
|
-
|
|
97
|
-
|
|
98
|
-
|
|
99
|
-
|
|
100
|
-
|
|
101
|
-
|
|
102
|
-
|
|
103
|
-
|
|
104
|
-
|
|
105
|
-
|
|
106
|
-
|
|
107
|
-
|
|
108
|
-
|
|
109
|
-
|
|
110
|
-
|
|
111
|
-
|
|
112
|
-
|
|
113
|
-
except Exception as e:
|
|
114
|
-
|
|
115
|
-
|
|
116
|
-
|
|
117
|
-
|
|
118
|
-
|
|
119
|
-
|
|
120
|
-
|
|
121
|
-
|
|
122
|
-
|
|
123
|
-
|
|
124
|
-
|
|
125
|
-
|
|
126
|
-
|
|
127
|
-
|
|
128
|
-
|
|
129
|
-
|
|
130
|
-
|
|
131
|
-
|
|
77
|
+
# eval_input_bias = [
|
|
78
|
+
# {
|
|
79
|
+
# "query": user_message,
|
|
80
|
+
# "context": context,
|
|
81
|
+
# "output": "", # No output yet
|
|
82
|
+
# }
|
|
83
|
+
# ]
|
|
84
|
+
# try:
|
|
85
|
+
# start_time = time.time()
|
|
86
|
+
# bias_evaluation_result = performEvaluation(
|
|
87
|
+
# eval_input_bias,
|
|
88
|
+
# api_key=self.llumo_key,
|
|
89
|
+
# evals=["Input Bias"],
|
|
90
|
+
# playgroundID=playground_id,
|
|
91
|
+
# workspaceID=workspace_id,
|
|
92
|
+
# createExperiment=create_experiment,
|
|
93
|
+
# )
|
|
94
|
+
# latency = int((time.time() - start_time) * 1000)
|
|
95
|
+
# # Access the first result object
|
|
96
|
+
# bias_evaluation = bias_evaluation_result[0]
|
|
97
|
+
# message = "-".join(
|
|
98
|
+
# getattr(bias_evaluation, "edgeCases", {}).get("value", [])
|
|
99
|
+
# )
|
|
100
|
+
# self.session.logEvalStep(
|
|
101
|
+
# stepName=f"EVAL-Input Bias",
|
|
102
|
+
# output="",
|
|
103
|
+
# context=context,
|
|
104
|
+
# query=user_message,
|
|
105
|
+
# messageHistory="",
|
|
106
|
+
# tools="",
|
|
107
|
+
# intermediateSteps="",
|
|
108
|
+
# groundTruth="",
|
|
109
|
+
# analyticsScore=getattr(bias_evaluation, "analyticsScore", {}),
|
|
110
|
+
# reasoning=getattr(bias_evaluation, "reasoning", {}),
|
|
111
|
+
# classification=getattr(bias_evaluation, "classification", {}),
|
|
112
|
+
# evalLabel=getattr(bias_evaluation, "evalLabel", {}),
|
|
113
|
+
# latencyMs=latency,
|
|
114
|
+
# status="SUCCESS",
|
|
115
|
+
# message=message,
|
|
116
|
+
# )
|
|
117
|
+
# except Exception as e:
|
|
118
|
+
# print(f"Input Bias evaluation failed: {e}")
|
|
119
|
+
# self.session.logEvalStep(
|
|
120
|
+
# stepName=f"EVAL-FAILURE",
|
|
121
|
+
# output="",
|
|
122
|
+
# context=context,
|
|
123
|
+
# query=user_message,
|
|
124
|
+
# messageHistory="",
|
|
125
|
+
# tools="",
|
|
126
|
+
# intermediateSteps="",
|
|
127
|
+
# groundTruth="",
|
|
128
|
+
# analyticsScore={},
|
|
129
|
+
# reasoning={},
|
|
130
|
+
# classification={},
|
|
131
|
+
# evalLabel={},
|
|
132
|
+
# latencyMs=0,
|
|
133
|
+
# status="FAILURE",
|
|
134
|
+
# message="EVAL_ERROR",
|
|
135
|
+
# )
|
|
132
136
|
|
|
133
137
|
start_time = time.time()
|
|
134
138
|
response = original_create(*args, **kwargs)
|
|
@@ -151,74 +155,77 @@ class OpenAI(OpenAIClient):
|
|
|
151
155
|
)
|
|
152
156
|
|
|
153
157
|
# Response Correctness Evaluation
|
|
154
|
-
eval_input_correctness = [
|
|
155
|
-
|
|
156
|
-
|
|
157
|
-
|
|
158
|
-
|
|
159
|
-
|
|
160
|
-
]
|
|
161
|
-
try:
|
|
162
|
-
|
|
163
|
-
|
|
164
|
-
|
|
165
|
-
|
|
166
|
-
|
|
167
|
-
|
|
168
|
-
|
|
169
|
-
|
|
170
|
-
|
|
171
|
-
|
|
172
|
-
|
|
173
|
-
|
|
174
|
-
|
|
175
|
-
|
|
176
|
-
|
|
177
|
-
|
|
178
|
-
|
|
179
|
-
|
|
180
|
-
|
|
181
|
-
|
|
182
|
-
|
|
183
|
-
|
|
184
|
-
|
|
185
|
-
|
|
186
|
-
|
|
187
|
-
|
|
188
|
-
|
|
189
|
-
|
|
190
|
-
|
|
191
|
-
|
|
192
|
-
|
|
193
|
-
|
|
194
|
-
|
|
195
|
-
|
|
196
|
-
|
|
197
|
-
|
|
198
|
-
except Exception as e:
|
|
199
|
-
|
|
200
|
-
|
|
201
|
-
|
|
202
|
-
|
|
203
|
-
|
|
204
|
-
|
|
205
|
-
|
|
206
|
-
|
|
207
|
-
|
|
208
|
-
|
|
209
|
-
|
|
210
|
-
|
|
211
|
-
|
|
212
|
-
|
|
213
|
-
|
|
214
|
-
|
|
215
|
-
|
|
216
|
-
|
|
217
|
-
|
|
218
|
-
|
|
219
|
-
if correctness_evaluation is None:
|
|
220
|
-
|
|
221
|
-
|
|
222
|
-
return ChatCompletionWithEval(
|
|
158
|
+
# eval_input_correctness = [
|
|
159
|
+
# {
|
|
160
|
+
# "query": user_message,
|
|
161
|
+
# "context": context,
|
|
162
|
+
# "output": output_text,
|
|
163
|
+
# }
|
|
164
|
+
# ]
|
|
165
|
+
# try:
|
|
166
|
+
# start_time = time.time()
|
|
167
|
+
# correctness_evaluation_result = performEvaluation(
|
|
168
|
+
# eval_input_correctness,
|
|
169
|
+
# api_key=self.llumo_key,
|
|
170
|
+
# evals=["Response Correctness"],
|
|
171
|
+
# playgroundID=playground_id,
|
|
172
|
+
# workspaceID=workspace_id,
|
|
173
|
+
# createExperiment=create_experiment,
|
|
174
|
+
# )
|
|
175
|
+
# latency = int((time.time() - start_time) * 1000)
|
|
176
|
+
# # Access the first result object
|
|
177
|
+
# correctness_evaluation = correctness_evaluation_result[0]
|
|
178
|
+
# message = "-".join(
|
|
179
|
+
# getattr(correctness_evaluation, "edgeCases", {}).get("value", [])
|
|
180
|
+
# )
|
|
181
|
+
# self.session.logEvalStep(
|
|
182
|
+
# stepName=f"EVAL-Response Correctness",
|
|
183
|
+
# output=output_text,
|
|
184
|
+
# context=context,
|
|
185
|
+
# query=user_message,
|
|
186
|
+
# messageHistory="",
|
|
187
|
+
# tools="",
|
|
188
|
+
# intermediateSteps="",
|
|
189
|
+
# groundTruth="",
|
|
190
|
+
# analyticsScore=getattr(
|
|
191
|
+
# correctness_evaluation, "analyticsScore", {}
|
|
192
|
+
# ),
|
|
193
|
+
# reasoning=getattr(correctness_evaluation, "reasoning", {}),
|
|
194
|
+
# classification=getattr(
|
|
195
|
+
# correctness_evaluation, "classification", {}
|
|
196
|
+
# ),
|
|
197
|
+
# evalLabel=getattr(correctness_evaluation, "evalLabel", {}),
|
|
198
|
+
# latencyMs=latency,
|
|
199
|
+
# status="SUCCESS",
|
|
200
|
+
# message=message,
|
|
201
|
+
# )
|
|
202
|
+
# except Exception as e:
|
|
203
|
+
# print(f"Response Correctness evaluation failed: {e}")
|
|
204
|
+
# correctness_evaluation = None
|
|
205
|
+
# self.session.logEvalStep(
|
|
206
|
+
# stepName=f"EVAL-FAILURE",
|
|
207
|
+
# output=output_text,
|
|
208
|
+
# context=context,
|
|
209
|
+
# query=user_message,
|
|
210
|
+
# messageHistory="",
|
|
211
|
+
# tools="",
|
|
212
|
+
# intermediateSteps="",
|
|
213
|
+
# groundTruth="",
|
|
214
|
+
# analyticsScore={},
|
|
215
|
+
# reasoning={},
|
|
216
|
+
# classification={},
|
|
217
|
+
# evalLabel={},
|
|
218
|
+
# latencyMs=0,
|
|
219
|
+
# status="FAILURE",
|
|
220
|
+
# message="EVAL_ERROR",
|
|
221
|
+
# )
|
|
222
|
+
|
|
223
|
+
# if correctness_evaluation is None:
|
|
224
|
+
# return response
|
|
225
|
+
|
|
226
|
+
return ChatCompletionWithEval(
|
|
227
|
+
response
|
|
228
|
+
# , correctness_evaluation
|
|
229
|
+
)
|
|
223
230
|
|
|
224
231
|
self.chat.completions.create = create_wrapper
|
|
@@ -6,15 +6,15 @@ llumo/client.py,sha256=keYx0GToNB-FXmGncXd0MOwwCGjxsIoDbOhTx2rCcMQ,71582
|
|
|
6
6
|
llumo/exceptions.py,sha256=1OyhN9YL9LcyUPUsqYHq6Rret0udATZAwMVJaio2_Ec,2123
|
|
7
7
|
llumo/execution.py,sha256=nWbJ7AvWuUPcOb6i-JzKRna_PvF-ewZTiK8skS-5n3w,1380
|
|
8
8
|
llumo/functionCalling.py,sha256=D5jYapu1rIvdIJNUYPYMTyhQ1H-6nkwoOLMi6eekfUE,7241
|
|
9
|
-
llumo/google.py,sha256=
|
|
9
|
+
llumo/google.py,sha256=6y9YnDFDRHv6-sQNT5LIsV9p31BCN0B9eow5KTRBWfM,2185
|
|
10
10
|
llumo/helpingFuntions.py,sha256=B6FwUQ5f1v4FKrWCbYoGWMFdscOV_liuuhTgNQ3cdrk,27275
|
|
11
11
|
llumo/llumoLogger.py,sha256=grdjhu6Ngxg7nhnrMOP5Pd5ALR7U2ROws48yhf_N7y0,1912
|
|
12
12
|
llumo/llumoSessionContext.py,sha256=iJDeC3FG6dID9jdaDK5d6hIlXWuhWx4KQy-Nnty4wYg,10485
|
|
13
13
|
llumo/models.py,sha256=aVEZsOOoQx5LeNtwSyBxqvrINq0izH3QWu_YjsMPE6o,2910
|
|
14
|
-
llumo/openai.py,sha256=
|
|
14
|
+
llumo/openai.py,sha256=PtcQhEEE5q5J7im9hd16FRP7h51VkNhFNX0txS2t7JY,8786
|
|
15
15
|
llumo/sockets.py,sha256=pBDo-U65hMIMwKMwZQl3iBkEjISEt-9BkXxZTWfSHF4,6116
|
|
16
|
-
llumo-0.2.
|
|
17
|
-
llumo-0.2.
|
|
18
|
-
llumo-0.2.
|
|
19
|
-
llumo-0.2.
|
|
20
|
-
llumo-0.2.
|
|
16
|
+
llumo-0.2.29.dist-info/licenses/LICENSE,sha256=tF9yAcfPV9xGT3ViWmC8hPvOo8BEk4ZICbUfcEo8Dlk,182
|
|
17
|
+
llumo-0.2.29.dist-info/METADATA,sha256=4h3zWbgU4QfCGBGtAlHvKom8hKOJrZwWnq44H8pic0I,1558
|
|
18
|
+
llumo-0.2.29.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
|
|
19
|
+
llumo-0.2.29.dist-info/top_level.txt,sha256=d5zUTMI99llPtLRB8rtSrqELm_bOqX-bNC5IcwlDk88,6
|
|
20
|
+
llumo-0.2.29.dist-info/RECORD,,
|
|
File without changes
|
|
File without changes
|
|
File without changes
|