mirror of
https://github.com/infiniflow/ragflow.git
synced 2025-12-08 20:42:30 +08:00
Rework logging (#3358)
Unified all log files into one. ### What problem does this PR solve? Unified all log files into one. ### Type of change - [x] Refactoring
This commit is contained in:
Zhichang Yu
@ -26,11 +26,12 @@ from api.db.db_models import Dialog, Conversation,DB
|
||||
from api.db.services.common_service import CommonService
|
||||
from api.db.services.knowledgebase_service import KnowledgebaseService
|
||||
from api.db.services.llm_service import LLMService, TenantLLMService, LLMBundle
|
||||
from api.settings import chat_logger, retrievaler, kg_retrievaler
|
||||
from api.settings import retrievaler, kg_retrievaler
|
||||
from rag.app.resume import forbidden_select_fields4resume
|
||||
from rag.nlp.search import index_name
|
||||
from rag.utils import rmSpace, num_tokens_from_string, encoder
|
||||
from api.utils.file_utils import get_project_base_directory
|
||||
from api.utils.log_utils import logger
|
||||
|
||||
|
||||
class DialogService(CommonService):
|
||||
@ -177,7 +178,7 @@ def chat(dialog, messages, stream=True, **kwargs):
|
||||
tts_mdl = LLMBundle(dialog.tenant_id, LLMType.TTS)
|
||||
# try to use sql if field mapping is good to go
|
||||
if field_map:
|
||||
chat_logger.info("Use SQL to retrieval:{}".format(questions[-1]))
|
||||
logger.info("Use SQL to retrieval:{}".format(questions[-1]))
|
||||
ans = use_sql(questions[-1], field_map, dialog.tenant_id, chat_mdl, prompt_config.get("quote", True))
|
||||
if ans:
|
||||
yield ans
|
||||
@ -219,7 +220,7 @@ def chat(dialog, messages, stream=True, **kwargs):
|
||||
doc_ids=attachments,
|
||||
top=dialog.top_k, aggs=False, rerank_mdl=rerank_mdl)
|
||||
knowledges = [ck["content_with_weight"] for ck in kbinfos["chunks"]]
|
||||
chat_logger.info(
|
||||
logger.info(
|
||||
"{}->{}".format(" ".join(questions), "\n->".join(knowledges)))
|
||||
retrieval_tm = timer()
|
||||
|
||||
@ -291,7 +292,7 @@ def chat(dialog, messages, stream=True, **kwargs):
|
||||
yield decorate_answer(answer)
|
||||
else:
|
||||
answer = chat_mdl.chat(prompt, msg[1:], gen_conf)
|
||||
chat_logger.info("User: {}|Assistant: {}".format(
|
||||
logger.info("User: {}|Assistant: {}".format(
|
||||
msg[-1]["content"], answer))
|
||||
res = decorate_answer(answer)
|
||||
res["audio_binary"] = tts(tts_mdl, answer)
|
||||
@ -319,8 +320,7 @@ def use_sql(question, field_map, tenant_id, chat_mdl, quota=True):
|
||||
nonlocal sys_prompt, user_promt, question, tried_times
|
||||
sql = chat_mdl.chat(sys_prompt, [{"role": "user", "content": user_promt}], {
|
||||
"temperature": 0.06})
|
||||
print(user_promt, sql)
|
||||
chat_logger.info(f"“{question}”==>{user_promt} get SQL: {sql}")
|
||||
logger.info(f"{question} ==> {user_promt} get SQL: {sql}")
|
||||
sql = re.sub(r"[\r\n]+", " ", sql.lower())
|
||||
sql = re.sub(r".*select ", "select ", sql.lower())
|
||||
sql = re.sub(r" +", " ", sql)
|
||||
@ -340,9 +340,7 @@ def use_sql(question, field_map, tenant_id, chat_mdl, quota=True):
|
||||
flds.append(k)
|
||||
sql = "select doc_id,docnm_kwd," + ",".join(flds) + sql[8:]
|
||||
|
||||
print(f"“{question}” get SQL(refined): {sql}")
|
||||
|
||||
chat_logger.info(f"“{question}” get SQL(refined): {sql}")
|
||||
logger.info(f"{question} get SQL(refined): {sql}")
|
||||
tried_times += 1
|
||||
return retrievaler.sql_retrieval(sql, format="json"), sql
|
||||
|
||||
@ -371,10 +369,9 @@ def use_sql(question, field_map, tenant_id, chat_mdl, quota=True):
|
||||
question, sql, tbl["error"]
|
||||
)
|
||||
tbl, sql = get_table()
|
||||
chat_logger.info("TRY it again: {}".format(sql))
|
||||
logger.info("TRY it again: {}".format(sql))
|
||||
|
||||
chat_logger.info("GET table: {}".format(tbl))
|
||||
print(tbl)
|
||||
logger.info("GET table: {}".format(tbl))
|
||||
if tbl.get("error") or len(tbl["rows"]) == 0:
|
||||
return None
|
||||
|
||||
@ -404,7 +401,7 @@ def use_sql(question, field_map, tenant_id, chat_mdl, quota=True):
|
||||
rows = re.sub(r"T[0-9]{2}:[0-9]{2}:[0-9]{2}(\.[0-9]+Z)?\|", "|", rows)
|
||||
|
||||
if not docid_idx or not docnm_idx:
|
||||
chat_logger.warning("SQL missing field: " + sql)
|
||||
logger.warning("SQL missing field: " + sql)
|
||||
return {
|
||||
"answer": "\n".join([clmns, line, rows]),
|
||||
"reference": {"chunks": [], "doc_aggs": []},
|
||||
|
||||
@ -17,7 +17,6 @@ import hashlib
|
||||
import json
|
||||
import random
|
||||
import re
|
||||
import traceback
|
||||
from concurrent.futures import ThreadPoolExecutor
|
||||
from copy import deepcopy
|
||||
from datetime import datetime
|
||||
@ -26,7 +25,7 @@ from io import BytesIO
|
||||
from peewee import fn
|
||||
|
||||
from api.db.db_utils import bulk_insert_into_db
|
||||
from api.settings import stat_logger, docStoreConn
|
||||
from api.settings import docStoreConn
|
||||
from api.utils import current_timestamp, get_format_time, get_uuid
|
||||
from graphrag.mind_map_extractor import MindMapExtractor
|
||||
from rag.settings import SVR_QUEUE_NAME
|
||||
@ -40,6 +39,7 @@ from api.db.services.common_service import CommonService
|
||||
from api.db.services.knowledgebase_service import KnowledgebaseService
|
||||
from api.db import StatusEnum
|
||||
from rag.utils.redis_conn import REDIS_CONN
|
||||
from api.utils.log_utils import logger
|
||||
|
||||
|
||||
class DocumentService(CommonService):
|
||||
@ -387,7 +387,7 @@ class DocumentService(CommonService):
|
||||
cls.update_by_id(d["id"], info)
|
||||
except Exception as e:
|
||||
if str(e).find("'0'") < 0:
|
||||
stat_logger.error("fetch task exception:" + str(e))
|
||||
logger.exception("fetch task exception")
|
||||
|
||||
@classmethod
|
||||
@DB.connection_context()
|
||||
@ -544,7 +544,7 @@ def doc_upload_and_parse(conversation_id, file_objs, user_id):
|
||||
"knowledge_graph_kwd": "mind_map"
|
||||
})
|
||||
except Exception as e:
|
||||
stat_logger.error("Mind map generation error:", traceback.format_exc())
|
||||
logger.exception("Mind map generation error")
|
||||
|
||||
vects = embedding(doc_id, [c["content_with_weight"] for c in cks])
|
||||
assert len(cks) == len(vects)
|
||||
|
||||
@ -28,6 +28,7 @@ from api.db.services.file2document_service import File2DocumentService
|
||||
from api.utils import get_uuid
|
||||
from api.utils.file_utils import filename_type, thumbnail_img
|
||||
from rag.utils.storage_factory import STORAGE_IMPL
|
||||
from api.utils.log_utils import logger
|
||||
|
||||
|
||||
class FileService(CommonService):
|
||||
@ -272,8 +273,8 @@ class FileService(CommonService):
|
||||
cls.delete_folder_by_pf_id(user_id, file.id)
|
||||
return cls.model.delete().where((cls.model.tenant_id == user_id)
|
||||
& (cls.model.id == folder_id)).execute(),
|
||||
except Exception as e:
|
||||
print(e)
|
||||
except Exception:
|
||||
logger.exception("delete_folder_by_pf_id")
|
||||
raise RuntimeError("Database error (File retrieval)!")
|
||||
|
||||
@classmethod
|
||||
@ -321,8 +322,8 @@ class FileService(CommonService):
|
||||
def move_file(cls, file_ids, folder_id):
|
||||
try:
|
||||
cls.filter_update((cls.model.id << file_ids, ), { 'parent_id': folder_id })
|
||||
except Exception as e:
|
||||
print(e)
|
||||
except Exception:
|
||||
logger.exception("move_file")
|
||||
raise RuntimeError("Database error (File move)!")
|
||||
|
||||
@classmethod
|
||||
|
||||
@ -14,12 +14,12 @@
|
||||
# limitations under the License.
|
||||
#
|
||||
from api.db.services.user_service import TenantService
|
||||
from api.settings import database_logger
|
||||
from rag.llm import EmbeddingModel, CvModel, ChatModel, RerankModel, Seq2txtModel, TTSModel
|
||||
from api.db import LLMType
|
||||
from api.db.db_models import DB
|
||||
from api.db.db_models import LLMFactories, LLM, TenantLLM
|
||||
from api.db.services.common_service import CommonService
|
||||
from api.utils.log_utils import logger
|
||||
|
||||
|
||||
class LLMFactoriesService(CommonService):
|
||||
@ -209,40 +209,40 @@ class LLMBundle(object):
|
||||
emd, used_tokens = self.mdl.encode(texts, batch_size)
|
||||
if not TenantLLMService.increase_usage(
|
||||
self.tenant_id, self.llm_type, used_tokens):
|
||||
database_logger.error(
|
||||
"Can't update token usage for {}/EMBEDDING used_tokens: {}".format(self.tenant_id, used_tokens))
|
||||
logger.error(
|
||||
"LLMBundle.encode can't update token usage for {}/EMBEDDING used_tokens: {}".format(self.tenant_id, used_tokens))
|
||||
return emd, used_tokens
|
||||
|
||||
def encode_queries(self, query: str):
|
||||
emd, used_tokens = self.mdl.encode_queries(query)
|
||||
if not TenantLLMService.increase_usage(
|
||||
self.tenant_id, self.llm_type, used_tokens):
|
||||
database_logger.error(
|
||||
"Can't update token usage for {}/EMBEDDING used_tokens: {}".format(self.tenant_id, used_tokens))
|
||||
logger.error(
|
||||
"LLMBundle.encode_queries can't update token usage for {}/EMBEDDING used_tokens: {}".format(self.tenant_id, used_tokens))
|
||||
return emd, used_tokens
|
||||
|
||||
def similarity(self, query: str, texts: list):
|
||||
sim, used_tokens = self.mdl.similarity(query, texts)
|
||||
if not TenantLLMService.increase_usage(
|
||||
self.tenant_id, self.llm_type, used_tokens):
|
||||
database_logger.error(
|
||||
"Can't update token usage for {}/RERANK used_tokens: {}".format(self.tenant_id, used_tokens))
|
||||
logger.error(
|
||||
"LLMBundle.similarity can't update token usage for {}/RERANK used_tokens: {}".format(self.tenant_id, used_tokens))
|
||||
return sim, used_tokens
|
||||
|
||||
def describe(self, image, max_tokens=300):
|
||||
txt, used_tokens = self.mdl.describe(image, max_tokens)
|
||||
if not TenantLLMService.increase_usage(
|
||||
self.tenant_id, self.llm_type, used_tokens):
|
||||
database_logger.error(
|
||||
"Can't update token usage for {}/IMAGE2TEXT used_tokens: {}".format(self.tenant_id, used_tokens))
|
||||
logger.error(
|
||||
"LLMBundle.describe can't update token usage for {}/IMAGE2TEXT used_tokens: {}".format(self.tenant_id, used_tokens))
|
||||
return txt
|
||||
|
||||
def transcription(self, audio):
|
||||
txt, used_tokens = self.mdl.transcription(audio)
|
||||
if not TenantLLMService.increase_usage(
|
||||
self.tenant_id, self.llm_type, used_tokens):
|
||||
database_logger.error(
|
||||
"Can't update token usage for {}/SEQUENCE2TXT used_tokens: {}".format(self.tenant_id, used_tokens))
|
||||
logger.error(
|
||||
"LLMBundle.transcription can't update token usage for {}/SEQUENCE2TXT used_tokens: {}".format(self.tenant_id, used_tokens))
|
||||
return txt
|
||||
|
||||
def tts(self, text):
|
||||
@ -250,8 +250,8 @@ class LLMBundle(object):
|
||||
if isinstance(chunk,int):
|
||||
if not TenantLLMService.increase_usage(
|
||||
self.tenant_id, self.llm_type, chunk, self.llm_name):
|
||||
database_logger.error(
|
||||
"Can't update token usage for {}/TTS".format(self.tenant_id))
|
||||
logger.error(
|
||||
"LLMBundle.tts can't update token usage for {}/TTS".format(self.tenant_id))
|
||||
return
|
||||
yield chunk
|
||||
|
||||
@ -259,8 +259,8 @@ class LLMBundle(object):
|
||||
txt, used_tokens = self.mdl.chat(system, history, gen_conf)
|
||||
if isinstance(txt, int) and not TenantLLMService.increase_usage(
|
||||
self.tenant_id, self.llm_type, used_tokens, self.llm_name):
|
||||
database_logger.error(
|
||||
"Can't update token usage for {}/CHAT llm_name: {}, used_tokens: {}".format(self.tenant_id, self.llm_name, used_tokens))
|
||||
logger.error(
|
||||
"LLMBundle.chat can't update token usage for {}/CHAT llm_name: {}, used_tokens: {}".format(self.tenant_id, self.llm_name, used_tokens))
|
||||
return txt
|
||||
|
||||
def chat_streamly(self, system, history, gen_conf):
|
||||
@ -268,7 +268,7 @@ class LLMBundle(object):
|
||||
if isinstance(txt, int):
|
||||
if not TenantLLMService.increase_usage(
|
||||
self.tenant_id, self.llm_type, txt, self.llm_name):
|
||||
database_logger.error(
|
||||
"Can't update token usage for {}/CHAT llm_name: {}, content: {}".format(self.tenant_id, self.llm_name, txt))
|
||||
logger.error(
|
||||
"LLMBundle.chat_streamly can't update token usage for {}/CHAT llm_name: {}, content: {}".format(self.tenant_id, self.llm_name, txt))
|
||||
return
|
||||
yield txt
|
||||
|
||||
Reference in New Issue
Block a user