add evaluate service api

api/controllers/service_api/app/completion.py

@@ -2,8 +2,11 @@ import json
 import logging
 from typing import Union, Generator
 
+from datasets import Dataset
 from flask import stream_with_context, Response
 from flask_restful import reqparse
+from ragas import evaluate
+from ragas.metrics import answer_relevancy, context_precision, faithfulness
 from werkzeug.exceptions import NotFound, InternalServerError
 
 import services
@@ -14,7 +17,8 @@ from controllers.service_api.app.error import AppUnavailableError, ProviderNotIn
     ProviderModelCurrentlyNotSupportError
 from controllers.service_api.wraps import AppApiResource
 from core.conversation_message_task import PubHandler
-from core.model_providers.error import LLMBadRequestError, LLMAuthorizationError, LLMAPIUnavailableError, LLMAPIConnectionError, \
+from core.model_providers.error import LLMBadRequestError, LLMAuthorizationError, LLMAPIUnavailableError, \
+    LLMAPIConnectionError, \
     LLMRateLimitError, ProviderTokenNotInitError, QuotaExceededError, ModelCurrentlyNotSupportError
 from libs.helper import uuid_value
 from services.completion_service import CompletionService
@@ -75,6 +79,84 @@ class CompletionApi(AppApiResource):
             raise InternalServerError()
 
 
+class CompletionEvaluateApi(AppApiResource):
+    def post(self, app_model, end_user):
+        if app_model.mode != 'chat':
+            raise AppUnavailableError()
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('inputs', type=dict, required=True, location='json')
+        parser.add_argument('queries', type=list, location='json', default='')
+        parser.add_argument('files', type=list, required=False, location='json')
+        parser.add_argument('response_mode', type=str, choices=['blocking', 'streaming'], location='json')
+        parser.add_argument('user', type=str, location='json')
+        parser.add_argument('retriever_from', type=str, required=False, default='dev', location='json')
+
+        args = parser.parse_args()
+
+        streaming = args['response_mode'] == 'streaming'
+
+        if end_user is None and args['user'] is not None:
+            end_user = create_or_update_end_user_for_user_id(app_model, args['user'])
+
+        args['auto_generate_name'] = False
+        queries = args['queries']
+        try:
+            questions = []
+            answers = []
+            contexts = []
+            for query in queries:
+                args['query'] = query
+                response = CompletionService.completion(
+                    app_model=app_model,
+                    user=end_user,
+                    args=args,
+                    from_source='api',
+                    streaming=streaming,
+                )
+                questions.append(query)
+                answers.append(response['answer'])
+                context = []
+                metadata = response['metadata']
+                if 'retriever_resources' in metadata and metadata['retriever_resources']:
+                    retriever_resources = response['metadata']['retriever_resources']
+                    for retriever_resource in retriever_resources:
+                        context.append(retriever_resource['content'])
+                else:
+                    context.append('')
+                contexts.append(context)
+            ds = Dataset.from_dict(
+                {
+                    "question": questions,
+                    "answer": answers,
+                    "contexts": contexts,
+                }
+            )
+            result = evaluate(ds, [answer_relevancy, context_precision, faithfulness])
+            return result
+        except services.errors.conversation.ConversationNotExistsError:
+            raise NotFound("Conversation Not Exists.")
+        except services.errors.conversation.ConversationCompletedError:
+            raise ConversationCompletedError()
+        except services.errors.app_model_config.AppModelConfigBrokenError:
+            logging.exception("App model config broken.")
+            raise AppUnavailableError()
+        except ProviderTokenNotInitError as ex:
+            raise ProviderNotInitializeError(ex.description)
+        except QuotaExceededError:
+            raise ProviderQuotaExceededError()
+        except ModelCurrentlyNotSupportError:
+            raise ProviderModelCurrentlyNotSupportError()
+        except (LLMBadRequestError, LLMAPIConnectionError, LLMAPIUnavailableError,
+                LLMRateLimitError, LLMAuthorizationError) as e:
+            raise CompletionRequestError(str(e))
+        except ValueError as e:
+            raise e
+        except Exception as e:
+            logging.exception("internal server error.")
+            raise InternalServerError()
+
+
 class CompletionStopApi(AppApiResource):
     def post(self, app_model, end_user, task_id):
         if app_model.mode != 'completion':
@@ -166,11 +248,13 @@ def compact_response(response: Union[dict | Generator]) -> Response:
                 logging.exception("App model config broken.")
                 yield "data: " + json.dumps(api.handle_error(AppUnavailableError()).get_json()) + "\n\n"
             except ProviderTokenNotInitError as ex:
-                yield "data: " + json.dumps(api.handle_error(ProviderNotInitializeError(ex.description)).get_json()) + "\n\n"
+                yield "data: " + json.dumps(
+                    api.handle_error(ProviderNotInitializeError(ex.description)).get_json()) + "\n\n"
             except QuotaExceededError:
                 yield "data: " + json.dumps(api.handle_error(ProviderQuotaExceededError()).get_json()) + "\n\n"
             except ModelCurrentlyNotSupportError:
-                yield "data: " + json.dumps(api.handle_error(ProviderModelCurrentlyNotSupportError()).get_json()) + "\n\n"
+                yield "data: " + json.dumps(
+                    api.handle_error(ProviderModelCurrentlyNotSupportError()).get_json()) + "\n\n"
             except (LLMBadRequestError, LLMAPIConnectionError, LLMAPIUnavailableError,
                     LLMRateLimitError, LLMAuthorizationError) as e:
                 yield "data: " + json.dumps(api.handle_error(CompletionRequestError(str(e))).get_json()) + "\n\n"
@@ -185,6 +269,7 @@ def compact_response(response: Union[dict | Generator]) -> Response:
 
 
 api.add_resource(CompletionApi, '/completion-messages')
+api.add_resource(CompletionEvaluateApi, '/chat-evaluate')
 api.add_resource(CompletionStopApi, '/completion-messages/<string:task_id>/stop')
 api.add_resource(ChatApi, '/chat-messages')
 api.add_resource(ChatStopApi, '/chat-messages/<string:task_id>/stop')

api/requirements.txt

@@ -54,4 +54,19 @@ zhipuai==1.0.7
 werkzeug==2.3.7
 pymilvus==2.3.0
 qdrant-client==1.6.4
-cohere~=4.32
\ No newline at end of file
+cohere~=4.32
+pydantic~=1.10.12
+requests~=2.31.0
+regex~=2023.8.8
+jieba3k~=0.35.1
+numpy~=1.25.2
+httpx~=0.24.1
+dataclasses~=0.6
+click~=8.1.6
+blinker~=1.6.2
+botocore~=1.31.17
+alembic~=1.11.2
+tqdm~=4.66.1
+pytz~=2022.7.1
+datasets~=2.15.0
+ragas~=0.0.20
\ No newline at end of file