init

app.py

@@ -0,0 +1,132 @@
+# Imports
+from langchain.document_loaders import PyPDFLoader
+import os
+from langchain.chains import RetrievalQA, ConversationalRetrievalChain
+from langchain.indexes import VectorstoreIndexCreator
+from langchain.text_splitter import CharacterTextSplitter, TokenTextSplitter
+from langchain.embeddings import OpenAIEmbeddings, HuggingFaceEmbeddings
+from langchain.vectorstores import Chroma
+from langchain import HuggingFacePipeline
+from langchain.chat_models import ChatOpenAI
+from dotenv import load_dotenv
+from langchain.memory import ConversationBufferMemory, ConversationTokenBufferMemory
+import gradio as gr
+
+# Funcion de carga de la api key
+def process_key(api_key):
+    os.environ['OPENAI_API_KEY'] = api_key
+
+def load_pdf(file):
+    name_file = file.name
+    print(file.name)
+    loader = PyPDFLoader(file.name)
+    documents = loader.load()
+    print(documents)
+    # Creo el objeto que permite dividir el texto en chunks
+    text_splitter = CharacterTextSplitter(chunk_size=1024, chunk_overlap=64)
+    # Esto lo que hace es dividir el texto en chunks de 2048 caracteres con un overlap de 128 caracteres
+    texts = text_splitter.split_documents(documents)
+    # Genero el objeto que crea los embeddings
+    # Nota: Estos embeddings son gratuitos a diferencia de los de OpenAI
+    embeddings = HuggingFaceEmbeddings()
+    # Defino el modelo de lenguaje
+    llm = ChatOpenAI(model='gpt-3.5-turbo', temperature=0.0, max_tokens=1000)
+    # Creo la base de datos de vectores
+    global vectorstore
+    vectorstore = Chroma.from_documents(texts, embeddings)
+    # Defino la memoria
+    
+    global memory
+    # La definicion de Memoria no es trivial, es bastante compleja de hecho se deben especificar bien todos los parameteros para que no de error
+    memory = ConversationTokenBufferMemory(llm=llm, 
+                                           memory_key="chat_history", 
+                                           input_key='question', 
+                                           output_key='answer', 
+                                           max_token_limit=1000,
+                                           return_messages=False)
+    # Defino la cadena de qa
+    global qa
+    qa = ConversationalRetrievalChain.from_llm(llm, 
+                                        vectorstore.as_retriever(search_kwargs={'k': 3}), # Este parametro especifica cuantos chunks se van a recuperar
+                                        return_source_documents=True,
+                                        verbose=True,
+                                        chain_type='stuff',
+                                        memory=memory,
+                                        max_tokens_limit=2500,
+                                        get_chat_history=lambda h: h)
+    return 'Done'
+
+# Funcion que ejecuta LLM y responde la pregunta
+def answer_question(question):
+    result = qa(inputs={'question': question})
+    pages = [x.metadata['page'] for i, x in enumerate(result['source_documents'])]
+    return result['answer'], pages
+
+# Funcion que pega las respuestas anteriores en el objeto Chat bot
+def bot(history):
+    res = qa(
+        {
+            'question': history[-1][0],
+            'chat_history': history[:-1]
+        }
+    )
+    history[-1][1] = res['answer']
+    return history
+
+# Agrego el texto a la historia del chat
+def add_text(history, text):
+    history = history + [(text, None)]
+    return history, ""
+
+# Analizar como parsea las ecuaciones
+with gr.Blocks() as demo:
+    with gr.Tab(label='Load PDF'):
+        with gr.Row():
+            with gr.Column():
+                open_ai_key = gr.Textbox(label='Ingresa tu api key de Open AI', type='password')
+        with gr.Row():
+            with gr.Column(scale=0.4):
+                api_key_button = gr.Button('Enviar', variant='primary')
+        with gr.Row():
+            pdf_file = gr.File(label='PDF file')
+            # Esta linea esta para probar si el calculo se realiza
+            emb = gr.Textbox(label='Calculo de Embeddings, por favor espere...')
+        # send_pdf = gr.Button(label='Load PDF').style(full_width=False)
+        with gr.Row():
+            with gr.Column(scale=0.50):
+                send_pdf = gr.Button(label='Load PDF')
+                send_pdf.click(load_pdf, pdf_file, emb)
+    with gr.Tab(label='Galicia QA Demo'):
+        chatbot = gr.Chatbot([], 
+                             elem_id="chatbot",
+                             label='Document GPT').style(height=500)
+        with gr.Row():
+            with gr.Column(scale=0.80):
+                txt = gr.Textbox(
+                    show_label=False,
+                    placeholder="Enter text and press enter",
+                ).style(container=False)
+
+            with gr.Column(scale=0.10):
+                submit_btn = gr.Button(
+                    'Submit',
+                    variant='primary'
+                )
+                
+            with gr.Column(scale=0.10):
+                clear_btn = gr.Button(
+                    'Clear',
+                    variant='stop'
+                )
+        # Tanto el submit (hacer enter en el campo de texto) como el submit_btn hacen la misma accion
+        txt.submit(fn=add_text, inputs=[chatbot, txt], outputs=[chatbot, txt] # Cuando envio el submit hago esta funcion
+                   ).then(fn=bot, inputs=chatbot, outputs=chatbot) # Luego hago esta otra funcion
+
+        submit_btn.click(fn=add_text, inputs=[chatbot, txt], outputs=[chatbot, txt]
+                         ).then(fn=bot, inputs=chatbot, outputs=chatbot)
+        
+        clear_btn.click(lambda: None, None, chatbot, queue=False)
+
+        api_key_button.click(fn=process_key, inputs=[open_ai_key], outputs=None)
+
+demo.launch(inline=False)

requirements.txt

@@ -0,0 +1,155 @@
+accelerate @ git+https://github.com/huggingface/accelerate.git@109f3272f542cbd0e34022f5455078a4ab99d7eb
+aiofiles==23.1.0
+aiohttp==3.8.4
+aiosignal==1.3.1
+altair==5.0.1
+anyio==3.7.0
+arxiv==1.4.7
+asttokens==2.2.1
+async-timeout==4.0.2
+attrs==23.1.0
+backcall==0.2.0
+backoff==2.2.1
+bitsandbytes @ https://github.com/acpopescu/bitsandbytes/releases/download/v0.38.0-win0/bitsandbytes-0.38.1-py3-none-any.whl
+certifi==2023.5.7
+charset-normalizer==3.1.0
+chromadb==0.3.25
+click==8.1.3
+clickhouse-connect==0.5.25
+colorama==0.4.6
+coloredlogs==15.0.1
+comm==0.1.3
+contourpy==1.0.7
+cycler==0.11.0
+dataclasses-json==0.5.7
+debugpy==1.6.7
+decorator==5.1.1
+diffusers==0.16.1
+duckdb==0.8.0
+einops==0.6.1
+exceptiongroup==1.1.1
+executing==1.2.0
+fastapi==0.96.0
+feedparser==6.0.10
+ffmpy==0.3.0
+filelock==3.12.0
+flatbuffers==23.5.26
+fonttools==4.39.4
+frozenlist==1.3.3
+fsspec==2023.5.0
+gradio==3.34.0
+gradio_client==0.2.6
+greenlet==2.0.2
+h11==0.14.0
+hnswlib==0.7.0
+httpcore==0.17.2
+httptools==0.5.0
+httpx==0.24.1
+huggingface-hub==0.15.1
+humanfriendly==10.0
+idna==3.4
+importlib-metadata==6.6.0
+ipykernel==6.23.1
+ipython==8.14.0
+ipywidgets==8.0.6
+jedi==0.18.2
+Jinja2==3.1.2
+joblib==1.2.0
+jsonschema==4.17.3
+jupyter_client==8.2.0
+jupyter_core==5.3.0
+jupyterlab-widgets==3.0.7
+kiwisolver==1.4.4
+langchain==0.0.189
+linkify-it-py==2.0.2
+lz4==4.3.2
+markdown-it-py==2.2.0
+MarkupSafe==2.1.2
+marshmallow==3.19.0
+marshmallow-enum==1.5.1
+matplotlib==3.7.1
+matplotlib-inline==0.1.6
+mdit-py-plugins==0.3.3
+mdurl==0.1.2
+monotonic==1.6
+mpmath==1.3.0
+multidict==6.0.4
+mypy-extensions==1.0.0
+nest-asyncio==1.5.6
+networkx==3.1
+nltk==3.8.1
+numexpr==2.8.4
+numpy==1.24.3
+onnxruntime==1.15.0
+openai==0.27.7
+openapi-schema-pydantic==1.2.4
+orjson==3.9.0
+overrides==7.3.1
+packaging==23.1
+pandas==2.0.2
+parso==0.8.3
+peft @ git+https://github.com/huggingface/peft.git@fcff23f005fc7bfb816ad1f55360442c170cd5f5
+pickleshare==0.7.5
+Pillow==9.3.0
+platformdirs==3.5.1
+posthog==3.0.1
+prompt-toolkit==3.0.38
+protobuf==4.23.2
+psutil==5.9.5
+pure-eval==0.2.2
+pydantic==1.10.8
+pydub==0.25.1
+Pygments==2.15.1
+pyodbc==4.0.39
+pyparsing==3.0.9
+pypdf==3.9.1
+pyreadline3==3.4.1
+pyrsistent==0.19.3
+python-dateutil==2.8.2
+python-dotenv==1.0.0
+python-multipart==0.0.6
+pytomlpp @ file:///D:/projects/proy__sql_langchain/pytomlpp-1.0.11-cp310-cp310-win_amd64.whl
+pytz==2023.3
+pywin32==306
+PyYAML==6.0
+pyzmq==25.1.0
+regex==2023.5.5
+requests==2.31.0
+safetensors==0.3.1
+scikit-learn==1.2.2
+scipy==1.10.1
+semantic-version==2.10.0
+sentence-transformers==2.2.2
+sentencepiece==0.1.99
+sgmllib3k==1.0.0
+six==1.16.0
+sniffio==1.3.0
+SQLAlchemy==2.0.15
+stack-data==0.6.2
+starlette==0.27.0
+sympy==1.12
+tenacity==8.2.2
+text-generation==0.6.0
+threadpoolctl==3.1.0
+tiktoken==0.4.0
+tokenizers==0.13.3
+toolz==0.12.0
+torch==2.0.1+cu117
+torchvision==0.15.2
+tornado==6.3.2
+tqdm==4.65.0
+traitlets==5.9.0
+transformers @ git+https://github.com/huggingface/transformers.git@bacaab1629972b85664fe61ec3caa4da7b55b041
+typing-inspect==0.9.0
+typing_extensions==4.6.3
+tzdata==2023.3
+uc-micro-py==1.0.2
+urllib3==2.0.2
+uvicorn==0.22.0
+watchfiles==0.19.0
+wcwidth==0.2.6
+websockets==11.0.3
+widgetsnbextension==4.0.7
+yarl==1.9.2
+zipp==3.15.0
+zstandard==0.21.0