Spaces:
Runtime error
Runtime error
Update main.py
Browse files
main.py
CHANGED
|
@@ -1,178 +1,77 @@
|
|
| 1 |
-
import
|
| 2 |
-
import
|
| 3 |
-
from
|
| 4 |
-
from
|
| 5 |
from langchain.schema.runnable import Runnable, RunnablePassthrough, RunnableLambda
|
| 6 |
from langchain.schema.runnable.config import RunnableConfig
|
| 7 |
-
from langchain.embeddings import HuggingFaceEmbeddings
|
| 8 |
-
from langchain.chains import ConversationalRetrievalChain
|
| 9 |
-
from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
|
| 10 |
-
from langchain.prompts import ChatPromptTemplate, MessagesPlaceholder
|
| 11 |
from langchain.schema import StrOutputParser
|
| 12 |
-
|
| 13 |
-
from langchain_pinecone import PineconeVectorStore
|
| 14 |
-
|
| 15 |
-
from pinecone import Pinecone
|
| 16 |
-
from langchain.memory import ChatMessageHistory, ConversationBufferMemory
|
| 17 |
import pandas as pd
|
| 18 |
import numpy as np
|
| 19 |
-
|
| 20 |
-
from
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
SystemMessagePromptTemplate,
|
| 24 |
-
)
|
| 25 |
-
from langchain_openai import ChatOpenAI
|
| 26 |
import chainlit as cl
|
| 27 |
-
from chainlit.input_widget import TextInput
|
| 28 |
-
from chainlit import user_session
|
| 29 |
-
from offres_emploi import Api
|
| 30 |
-
from offres_emploi.utils import dt_to_str_iso
|
| 31 |
-
import datetime
|
| 32 |
-
import bcrypt
|
| 33 |
-
import json
|
| 34 |
|
| 35 |
-
|
| 36 |
-
def auth_callback(username: str, password: str):
|
| 37 |
-
auth = json.loads(os.environ['CHAINLIT_AUTH_LOGIN'])
|
| 38 |
-
ident = next(d['ident'] for d in auth if d['ident'] == username)
|
| 39 |
-
pwd = next(d['pwd'] for d in auth if d['ident'] == username)
|
| 40 |
-
resultLogAdmin = bcrypt.checkpw(username.encode('utf-8'), bcrypt.hashpw(ident.encode('utf-8'), bcrypt.gensalt()))
|
| 41 |
-
resultPwdAdmin = bcrypt.checkpw(password.encode('utf-8'), bcrypt.hashpw(pwd.encode('utf-8'), bcrypt.gensalt()))
|
| 42 |
-
resultRole = next(d['role'] for d in auth if d['ident'] == username)
|
| 43 |
-
if resultLogAdmin and resultPwdAdmin and resultRole == "admindatapcc":
|
| 44 |
-
return cl.User(
|
| 45 |
-
identifier=ident + " : 🧑💼 Admin Datapcc", metadata={"role": "admin", "provider": "credentials"}
|
| 46 |
-
)
|
| 47 |
-
elif resultLogAdmin and resultPwdAdmin and resultRole == "userdatapcc":
|
| 48 |
-
return cl.User(
|
| 49 |
-
identifier=ident + " : 🧑🎓 User Datapcc", metadata={"role": "user", "provider": "credentials"}
|
| 50 |
-
)
|
| 51 |
-
|
| 52 |
-
os.environ["TOKENIZERS_PARALLELISM"] = os.environ["TOKENIZERS_PARALLELISM"]
|
| 53 |
-
os.environ['OPENAI_API_KEY'] = os.environ['OPENAI_API_KEY']
|
| 54 |
|
|
|
|
|
|
|
| 55 |
@cl.author_rename
|
| 56 |
def rename(orig_author: str):
|
| 57 |
-
rename_dict = {"
|
| 58 |
return rename_dict.get(orig_author, orig_author)
|
| 59 |
|
| 60 |
-
@cl.action_callback("download")
|
| 61 |
-
async def on_action(action):
|
| 62 |
-
content = []
|
| 63 |
-
content.append(action.value)
|
| 64 |
-
arrayContent = np.array(content)
|
| 65 |
-
df = pd.DataFrame(arrayContent)
|
| 66 |
-
with open('./' + action.description + '.txt', 'wb') as csv_file:
|
| 67 |
-
df.to_csv(path_or_buf=csv_file, index=False,header=False, encoding='utf-8')
|
| 68 |
-
elements = [
|
| 69 |
-
cl.File(
|
| 70 |
-
name= action.description + ".txt",
|
| 71 |
-
path="./" + action.description + ".txt",
|
| 72 |
-
display="inline",
|
| 73 |
-
),
|
| 74 |
-
]
|
| 75 |
-
await cl.Message(
|
| 76 |
-
author="Datapcc 🌐🌐🌐", content="[Lien] 🔗", elements=elements
|
| 77 |
-
).send()
|
| 78 |
-
await action.remove()
|
| 79 |
-
|
| 80 |
-
def retriever_to_cache():
|
| 81 |
-
os.environ['PINECONE_API_KEY'] = os.environ['PINECONE_API_KEY']
|
| 82 |
-
os.environ['PINECONE_ENVIRONMENT'] = "us-west4-gcp-free"
|
| 83 |
-
index_name = os.environ['PINECONE_INDEX_NAME']
|
| 84 |
-
embeddings = HuggingFaceEmbeddings()
|
| 85 |
-
vectorstore = PineconeVectorStore(
|
| 86 |
-
index_name=index_name, embedding=embeddings
|
| 87 |
-
)
|
| 88 |
-
retriever = vectorstore.as_retriever(search_type="similarity_score_threshold", search_kwargs={"score_threshold": .7, "k": 30,"filter": {'categorie': {'$eq': 'OF'}}})
|
| 89 |
-
return retriever
|
| 90 |
-
|
| 91 |
-
@cl.set_chat_profiles
|
| 92 |
-
async def chat_profile():
|
| 93 |
-
return [
|
| 94 |
-
cl.ChatProfile(name="OF - Offre de formation",markdown_description="Requêter sur l'offre de formation - OF",icon="./public/favicon.png",),
|
| 95 |
-
]
|
| 96 |
@cl.on_chat_start
|
| 97 |
-
async def
|
| 98 |
-
|
| 99 |
-
|
| 100 |
-
|
| 101 |
-
|
| 102 |
-
|
| 103 |
-
|
| 104 |
-
|
| 105 |
-
|
| 106 |
-
|
| 107 |
-
|
| 108 |
-
|
| 109 |
-
).send()
|
| 110 |
|
| 111 |
-
|
| 112 |
-
|
| 113 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 114 |
|
| 115 |
-
Chat History:
|
| 116 |
-
{chat_history}
|
| 117 |
-
Follow Up Input: {question}
|
| 118 |
-
Standalone question:"""
|
| 119 |
-
CONDENSE_QUESTION_PROMPT = PromptTemplate.from_template(_template)
|
| 120 |
-
|
| 121 |
-
########## Chain with streaming ##########
|
| 122 |
-
message_history = ChatMessageHistory()
|
| 123 |
-
memory = ConversationBufferMemory(
|
| 124 |
-
memory_key="chat_history",
|
| 125 |
-
output_key="answer",
|
| 126 |
-
chat_memory=message_history,
|
| 127 |
-
return_messages=True,
|
| 128 |
-
)
|
| 129 |
-
streaming_llm = ChatOpenAI(
|
| 130 |
-
model_name = "gpt-4-1106-preview",
|
| 131 |
-
streaming=True,
|
| 132 |
-
temperature=1
|
| 133 |
-
)
|
| 134 |
-
qa = ConversationalRetrievalChain.from_llm(
|
| 135 |
-
streaming_llm,
|
| 136 |
-
memory=memory,
|
| 137 |
-
chain_type="stuff",
|
| 138 |
-
return_source_documents=True,
|
| 139 |
-
verbose=False,
|
| 140 |
-
retriever=retriever_to_cache()
|
| 141 |
-
)
|
| 142 |
-
cl.user_session.set("conversation_chain", qa)
|
| 143 |
-
|
| 144 |
@cl.on_message
|
| 145 |
-
async def
|
| 146 |
-
|
| 147 |
-
|
| 148 |
-
|
| 149 |
-
|
| 150 |
-
|
| 151 |
-
|
| 152 |
-
|
| 153 |
-
|
| 154 |
-
|
| 155 |
-
|
| 156 |
-
|
| 157 |
-
|
| 158 |
-
|
| 159 |
-
|
| 160 |
-
|
| 161 |
-
|
| 162 |
-
|
| 163 |
-
|
| 164 |
-
|
| 165 |
-
|
| 166 |
-
|
| 167 |
-
|
| 168 |
-
|
| 169 |
-
|
| 170 |
-
|
| 171 |
-
cl.
|
| 172 |
-
|
| 173 |
-
|
| 174 |
-
await cl.Message(author="Datapcc 🌐🌐🌐",content=answer).send()
|
| 175 |
-
await cl.Message(author="Datapcc 🌐🌐🌐",content="Download", actions=actions).send()
|
| 176 |
-
|
| 177 |
-
if metadatas:
|
| 178 |
-
await cl.Message(author="Datapcc 🌐🌐🌐",content="Sources : " + metadatas, elements=text_elements).send()
|
|
|
|
| 1 |
+
from langchain_anthropic import ChatAnthropic
|
| 2 |
+
from langchain.prompts import ChatPromptTemplate, MessagesPlaceholder
|
| 3 |
+
from langchain.memory import ChatMessageHistory, ConversationBufferMemory
|
| 4 |
+
from langchain.schema import StrOutputParser
|
| 5 |
from langchain.schema.runnable import Runnable, RunnablePassthrough, RunnableLambda
|
| 6 |
from langchain.schema.runnable.config import RunnableConfig
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
from langchain.schema import StrOutputParser
|
| 8 |
+
import os
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
import pandas as pd
|
| 10 |
import numpy as np
|
| 11 |
+
|
| 12 |
+
from langchain.agents.agent_types import AgentType
|
| 13 |
+
from langchain_experimental.agents.agent_toolkits import create_csv_agent
|
| 14 |
+
|
|
|
|
|
|
|
|
|
|
| 15 |
import chainlit as cl
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
|
| 17 |
+
os.environ["ANTHROPIC_API_KEY"] = os.environ["ANTHROPIC_API_KEY"]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
|
| 19 |
+
def library():
|
| 20 |
+
return "Exemple de requêtes sur les données de l'enquête.\n\nQ1 : Quels sont les équipements préférentiels des étudiant.e.s?\nQ2 : Quels sont les 3 outils numériques principaux de l'université pour le travail universitaire?\nQ3 : Quels sont les outils numériques de l'université préférés des étudiant.e.s?\nQ4 : Quels sont les réseaux sociaux préférés des étudiant.e.s?\nQ5 : Quels sont les outils numériques de l'université préférés des étudiant.e.s pour communiquer?\nQ6 : Quels sont les outils numériques de l'université préférés des étudiant.e.s pour le travail universitaire?\nQ7 : Quel est l'usage du mail de l'université?\nQ8 : Quel est l'usage de l'ENT de l'université?\nQ9 : Donne le pourcentage d'étudiant.e.s en licence3 qui utilise souvent Moodle?\nQ10 : Donne le pourcentage d'étudiant.e.s en licence1 qui utilise souvent le mail?\nQ11 : Donne le pourcentage d'étudiant.e.s en licence1 de la filière Sciences économiques qui utilise souvent le mail?\nQ12 : Pourquoi les étudiants utilisent WhatsApp?\nQ13 : Pourquoi les étudiants utilisent Discord?\nQ14 : Quels avantages représentent les outils numériques?\nQ15 : Quelles sont les principales difficultés?"
|
| 21 |
@cl.author_rename
|
| 22 |
def rename(orig_author: str):
|
| 23 |
+
rename_dict = {"AgentExecutor": "Agent conversationnel", "Error": "Réponse de l'assistant", "Datapcc Chain": "Copilot", "load_memory_variables": "Historique de conversation 💬", "Retriever": "Agent conversationnel", "StuffDocumentsChain": "Chaîne de documents", "LLMChain": "Agent", "ChatAnthropic": "Réponse de l'IA 🤖"}
|
| 24 |
return rename_dict.get(orig_author, orig_author)
|
| 25 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
@cl.on_chat_start
|
| 27 |
+
async def on_chat_start():
|
| 28 |
+
await cl.Message(f"> Votre assistant conversationnel vous permet d'analyser les données de l'enquête sur les usages du numérique par les étudiant.e.s Licence").send()
|
| 29 |
+
listPrompts_name = f"Liste des requêtes"
|
| 30 |
+
prompt_elements = []
|
| 31 |
+
prompt_elements.append(
|
| 32 |
+
cl.Text(content=library(), name=listPrompts_name)
|
| 33 |
+
)
|
| 34 |
+
await cl.Message(content="📚 Bibliothèque de questions : " + listPrompts_name, elements=prompt_elements).send()
|
| 35 |
+
await cl.Avatar(
|
| 36 |
+
name="You",
|
| 37 |
+
path="./logo-ofipe.jpg",
|
| 38 |
+
).send()
|
|
|
|
| 39 |
|
| 40 |
+
agent = create_csv_agent(
|
| 41 |
+
ChatAnthropic(temperature=1,model_name="claude-3-sonnet-20240229"),
|
| 42 |
+
"./enquete.csv",
|
| 43 |
+
verbose=False,
|
| 44 |
+
agent_type=AgentType.ZERO_SHOT_REACT_DESCRIPTION,
|
| 45 |
+
)
|
| 46 |
+
cl.user_session.set("runnable", agent)
|
| 47 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 48 |
@cl.on_message
|
| 49 |
+
async def on_message(message: cl.Message):
|
| 50 |
+
memory= cl.user_session.get("memory")
|
| 51 |
+
runnable = cl.user_session.get("runnable") # type: Runnable
|
| 52 |
+
cb = cl.AsyncLangchainCallbackHandler()
|
| 53 |
+
try:
|
| 54 |
+
res = await runnable.acall("Réponds en langue française à la question suivante :\n" + message.content + "\nDétaille la réponse en faisant une analyse complète.", callbacks=[cb])
|
| 55 |
+
await cl.Message(author="COPILOT",content=res['output']).send()
|
| 56 |
+
listPrompts_name = f"Liste des requêtes"
|
| 57 |
+
prompt_elements = []
|
| 58 |
+
prompt_elements.append(
|
| 59 |
+
cl.Text(content=library(), name=listPrompts_name)
|
| 60 |
+
)
|
| 61 |
+
await cl.Message(content="📚 Bibliothèque de questions : " + listPrompts_name, elements=prompt_elements).send()
|
| 62 |
+
except ValueError as e:
|
| 63 |
+
res = str(e)
|
| 64 |
+
resArray = res.split(":")
|
| 65 |
+
ans = ''
|
| 66 |
+
if str(res).find('parsing') != -1:
|
| 67 |
+
for i in range(2,len(resArray)):
|
| 68 |
+
ans += resArray[i]
|
| 69 |
+
await cl.Message(author="COPILOT",content=ans.replace("`","")).send()
|
| 70 |
+
listPrompts_name = f"Liste des requêtes"
|
| 71 |
+
prompt_elements = []
|
| 72 |
+
prompt_elements.append(
|
| 73 |
+
cl.Text(content=library(), name=listPrompts_name)
|
| 74 |
+
)
|
| 75 |
+
await cl.Message(content="📚 Bibliothèque de questions : " + listPrompts_name, elements=prompt_elements).send()
|
| 76 |
+
else:
|
| 77 |
+
await cl.Message(author="COPILOT",content="Reformulez votre requête, s'il vous plait 😃").send()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|