Initial commit

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.sqlite3 filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1 @@
+.venv

.python-version

@@ -0,0 +1 @@
+3.10.10

README.md

@@ -4,7 +4,7 @@ emoji: 🏢
 colorFrom: pink
 colorTo: gray
 sdk: gradio
-sdk_version: 4.31.2
+sdk_version: 4.31.0
 app_file: app.py
 pinned: false
 ---

app.py

@@ -0,0 +1,88 @@
+__import__('pysqlite3')
+import sys
+sys.modules['sqlite3'] = sys.modules.pop('pysqlite3')
+
+from langchain_chroma import Chroma
+from langchain_openai import OpenAIEmbeddings
+from langchain_core.messages import HumanMessage
+from langchain.chains import create_history_aware_retriever, create_retrieval_chain
+from langchain.chains.combine_documents import create_stuff_documents_chain
+from langchain_chroma import Chroma
+from langchain_openai import OpenAIEmbeddings
+from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder, PromptTemplate
+from langchain_openai import ChatOpenAI
+from dotenv import load_dotenv
+import gradio as gr
+
+load_dotenv()
+
+# OpenAI's GPT model to use
+GPT_VERSION = "gpt-3.5-turbo-0125"
+
+# Prompt used to contextualize a question inside the context of a conversation
+CONTEXTUALIZATION_PROMPT = """Given a chat history and the latest user question \
+which might reference context in the chat history, formulate a standalone question \
+which can be understood without the chat history. Do NOT answer the question, \
+just reformulate it if needed and otherwise return it as is."""
+
+# Prompt used for building an answer to a question using the retrieved context
+SYSTEM_PROMPT = """You are a marketing assistant for question-answering tasks for a company called Tryolabs. \
+You must answer the question using pieces of retrieved context consisting of blogposts written by Tryolabs. \
+Answer the question in a friendly manner, trying to refer to Tryolabs' content to engage the user and encourage \
+to make contact with the company. \
+When you can, include properly formatted hyperlinks to the content. TRY providing references to different content \
+within the same response. Extract the hyperlink from the context. \
+If you don't know the answer to the question, encourage the user to contact us for more information. \
+Tryolabs' contact email is hello@tryolabs.com. \
+Keep your answers concise. \
+ALWAYS respond with Markdown. \
+You MUST only use information found on the pieces of context given. You MUST NOT attribute any inventions to \
+Tryolabs unless explicitly mentioned on the context. \
+
+{context}"""
+
+llm = ChatOpenAI(model=GPT_VERSION)
+
+vectorstore = Chroma(persist_directory="./chroma_db", embedding_function=OpenAIEmbeddings())
+
+retriever = vectorstore.as_retriever(search_type="mmr", search_kwargs={'k': 6, 'lambda_mult': 0.3})
+
+contextualize_q_prompt = ChatPromptTemplate.from_messages(
+    [
+        ("system", CONTEXTUALIZATION_PROMPT),
+        MessagesPlaceholder("chat_history"),
+        ("human", "{input}"),
+    ]
+)
+
+history_aware_retriever = create_history_aware_retriever(llm, retriever, contextualize_q_prompt)
+
+custom_rag_prompt = ChatPromptTemplate.from_messages(
+    [
+        ("system", SYSTEM_PROMPT),
+        MessagesPlaceholder("chat_history"),
+        ("human", "{input}")
+    ]
+)
+
+document_prompt = PromptTemplate(
+    input_variables=['page_content', 'title', 'link'], 
+    template="Title: {title}, Link to blogpost: {link}, Content: {page_content}"
+)
+qa_chain = create_stuff_documents_chain(llm, custom_rag_prompt, document_prompt=document_prompt)
+rag_chain = create_retrieval_chain(history_aware_retriever, qa_chain)
+
+
+def predict(message, history):
+    chat_history = []
+
+    for m, a in history:
+        chat_history.append(HumanMessage(content=m))
+        chat_history.append(a)
+
+    ai_response = rag_chain.invoke({"input": message, "chat_history": chat_history})
+
+    return ai_response["answer"]
+
+
+gr.ChatInterface(predict).launch()

chroma_db/40cfeb6b-ca96-4457-986f-a2c4b1cb67b0/data_level0.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e80579c3770fa3828f94929cfac1344518de8f9fc10cac262ee6939f28e36df1
+size 6284000

chroma_db/40cfeb6b-ca96-4457-986f-a2c4b1cb67b0/header.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b79d047db53c51e35116714da0b758090f648594d1ad044284bdd6f6756eb567
+size 100

chroma_db/40cfeb6b-ca96-4457-986f-a2c4b1cb67b0/index_metadata.pickle

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b9e84ff4ca6cbcf856227323fc2e32d7275fde24dc3d55941ccc01a5910674da
+size 55974

chroma_db/40cfeb6b-ca96-4457-986f-a2c4b1cb67b0/length.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:581553af82e8aac6ab852c7c08b1347250ce2ca32abdbe206ad2234689ff4636
+size 4000

chroma_db/40cfeb6b-ca96-4457-986f-a2c4b1cb67b0/link_lists.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ef528985e10e55fc2d1b32cc50a14d0845772f6e3cbd584fa032df651dc1945
+size 8420

chroma_db/chroma.sqlite3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:810ce85ed0c03b6ad2c3879aad9724abc31b1bb2c43478d1dde2bb26e5dc4112
+size 16801792

requirements.txt

@@ -0,0 +1,199 @@
+aiofiles==23.2.1 ; python_full_version == "3.10.10"
+aiohttp==3.9.5 ; python_full_version == "3.10.10"
+aiosignal==1.3.1 ; python_full_version == "3.10.10"
+altair==5.3.0 ; python_full_version == "3.10.10"
+annotated-types==0.6.0 ; python_full_version == "3.10.10"
+anyio==4.3.0 ; python_full_version == "3.10.10"
+appnope==0.1.4 ; platform_system == "Darwin" and python_full_version == "3.10.10"
+asgiref==3.8.1 ; python_full_version == "3.10.10"
+asttokens==2.4.1 ; python_full_version == "3.10.10"
+async-timeout==4.0.3 ; python_full_version == "3.10.10"
+attrs==23.2.0 ; python_full_version == "3.10.10"
+backoff==2.2.1 ; python_full_version == "3.10.10"
+bcrypt==4.1.3 ; python_full_version == "3.10.10"
+beautifulsoup4==4.12.3 ; python_full_version == "3.10.10"
+bs4==0.0.2 ; python_full_version == "3.10.10"
+build==1.2.1 ; python_full_version == "3.10.10"
+cachetools==5.3.3 ; python_full_version == "3.10.10"
+certifi==2024.2.2 ; python_full_version == "3.10.10"
+cffi==1.16.0 ; implementation_name == "pypy" and python_full_version == "3.10.10"
+charset-normalizer==3.3.2 ; python_full_version == "3.10.10"
+chroma-hnswlib==0.7.3 ; python_full_version == "3.10.10"
+chromadb==0.4.24 ; python_full_version == "3.10.10"
+click==8.1.7 ; python_full_version == "3.10.10"
+colorama==0.4.6 ; sys_platform == "win32" and python_full_version == "3.10.10" or platform_system == "Windows" and python_full_version == "3.10.10" or os_name == "nt" and python_full_version == "3.10.10"
+coloredlogs==15.0.1 ; python_full_version == "3.10.10"
+comm==0.2.2 ; python_full_version == "3.10.10"
+contourpy==1.2.1 ; python_full_version == "3.10.10"
+cssselect==1.2.0 ; python_full_version == "3.10.10"
+cycler==0.12.1 ; python_full_version == "3.10.10"
+dataclasses-json==0.6.6 ; python_full_version == "3.10.10"
+debugpy==1.8.1 ; python_full_version == "3.10.10"
+decorator==5.1.1 ; python_full_version == "3.10.10"
+deprecated==1.2.14 ; python_full_version == "3.10.10"
+distro==1.9.0 ; python_full_version == "3.10.10"
+dnspython==2.6.1 ; python_full_version == "3.10.10"
+email-validator==2.1.1 ; python_full_version == "3.10.10"
+exceptiongroup==1.2.1 ; python_full_version == "3.10.10"
+executing==2.0.1 ; python_full_version == "3.10.10"
+fastapi-cli==0.0.3 ; python_full_version == "3.10.10"
+fastapi==0.111.0 ; python_full_version == "3.10.10"
+feedfinder2==0.0.4 ; python_full_version == "3.10.10"
+feedparser==6.0.11 ; python_full_version == "3.10.10"
+ffmpy==0.3.2 ; python_full_version == "3.10.10"
+filelock==3.14.0 ; python_full_version == "3.10.10"
+flatbuffers==24.3.25 ; python_full_version == "3.10.10"
+fonttools==4.51.0 ; python_full_version == "3.10.10"
+frozenlist==1.4.1 ; python_full_version == "3.10.10"
+fsspec==2024.3.1 ; python_full_version == "3.10.10"
+google-auth==2.29.0 ; python_full_version == "3.10.10"
+googleapis-common-protos==1.63.0 ; python_full_version == "3.10.10"
+gradio-client==0.16.2 ; python_full_version == "3.10.10"
+gradio==4.31.0 ; python_full_version == "3.10.10"
+greenlet==3.0.3 ; platform_machine == "aarch64" and python_full_version == "3.10.10" or platform_machine == "ppc64le" and python_full_version == "3.10.10" or platform_machine == "x86_64" and python_full_version == "3.10.10" or platform_machine == "amd64" and python_full_version == "3.10.10" or platform_machine == "AMD64" and python_full_version == "3.10.10" or platform_machine == "win32" and python_full_version == "3.10.10" or platform_machine == "WIN32" and python_full_version == "3.10.10"
+grpcio==1.63.0 ; python_full_version == "3.10.10"
+h11==0.14.0 ; python_full_version == "3.10.10"
+httpcore==1.0.5 ; python_full_version == "3.10.10"
+httptools==0.6.1 ; python_full_version == "3.10.10"
+httpx==0.27.0 ; python_full_version == "3.10.10"
+huggingface-hub==0.23.0 ; python_full_version == "3.10.10"
+humanfriendly==10.0 ; python_full_version == "3.10.10"
+idna==3.7 ; python_full_version == "3.10.10"
+importlib-metadata==7.0.0 ; python_full_version == "3.10.10"
+importlib-resources==6.4.0 ; python_full_version == "3.10.10"
+ipdb==0.13.13 ; python_full_version == "3.10.10"
+ipykernel==6.29.4 ; python_full_version == "3.10.10"
+ipython==8.24.0 ; python_full_version == "3.10.10"
+jedi==0.19.1 ; python_full_version == "3.10.10"
+jieba3k==0.35.1 ; python_full_version == "3.10.10"
+jinja2==3.1.4 ; python_full_version == "3.10.10"
+joblib==1.4.2 ; python_full_version == "3.10.10"
+jsonpatch==1.33 ; python_full_version == "3.10.10"
+jsonpointer==2.4 ; python_full_version == "3.10.10"
+jsonschema-specifications==2023.12.1 ; python_full_version == "3.10.10"
+jsonschema==4.22.0 ; python_full_version == "3.10.10"
+jupyter-client==8.6.1 ; python_full_version == "3.10.10"
+jupyter-core==5.7.2 ; python_full_version == "3.10.10"
+kiwisolver==1.4.5 ; python_full_version == "3.10.10"
+kubernetes==29.0.0 ; python_full_version == "3.10.10"
+langchain-chroma==0.1.0 ; python_full_version == "3.10.10"
+langchain-community==0.0.38 ; python_full_version == "3.10.10"
+langchain-core==0.1.52 ; python_full_version == "3.10.10"
+langchain-openai==0.1.6 ; python_full_version == "3.10.10"
+langchain-text-splitters==0.0.1 ; python_full_version == "3.10.10"
+langchain==0.1.20 ; python_full_version == "3.10.10"
+langchainhub==0.1.15 ; python_full_version == "3.10.10"
+langsmith==0.1.57 ; python_full_version == "3.10.10"
+lxml-html-clean==0.1.1 ; python_full_version == "3.10.10"
+lxml==5.2.2 ; python_full_version == "3.10.10"
+markdown-it-py==3.0.0 ; python_full_version == "3.10.10"
+markupsafe==2.1.5 ; python_full_version == "3.10.10"
+marshmallow==3.21.2 ; python_full_version == "3.10.10"
+matplotlib-inline==0.1.7 ; python_full_version == "3.10.10"
+matplotlib==3.8.4 ; python_full_version == "3.10.10"
+mdurl==0.1.2 ; python_full_version == "3.10.10"
+mmh3==4.1.0 ; python_full_version == "3.10.10"
+monotonic==1.6 ; python_full_version == "3.10.10"
+mpmath==1.3.0 ; python_full_version == "3.10.10"
+multidict==6.0.5 ; python_full_version == "3.10.10"
+mypy-extensions==1.0.0 ; python_full_version == "3.10.10"
+nest-asyncio==1.6.0 ; python_full_version == "3.10.10"
+newspaper3k==0.2.8 ; python_full_version == "3.10.10"
+nltk==3.8.1 ; python_full_version == "3.10.10"
+numpy==1.26.4 ; python_full_version == "3.10.10"
+oauthlib==3.2.2 ; python_full_version == "3.10.10"
+onnxruntime==1.17.3 ; python_full_version == "3.10.10"
+openai==1.28.1 ; python_full_version == "3.10.10"
+opentelemetry-api==1.24.0 ; python_full_version == "3.10.10"
+opentelemetry-exporter-otlp-proto-common==1.24.0 ; python_full_version == "3.10.10"
+opentelemetry-exporter-otlp-proto-grpc==1.24.0 ; python_full_version == "3.10.10"
+opentelemetry-instrumentation-asgi==0.45b0 ; python_full_version == "3.10.10"
+opentelemetry-instrumentation-fastapi==0.45b0 ; python_full_version == "3.10.10"
+opentelemetry-instrumentation==0.45b0 ; python_full_version == "3.10.10"
+opentelemetry-proto==1.24.0 ; python_full_version == "3.10.10"
+opentelemetry-sdk==1.24.0 ; python_full_version == "3.10.10"
+opentelemetry-semantic-conventions==0.45b0 ; python_full_version == "3.10.10"
+opentelemetry-util-http==0.45b0 ; python_full_version == "3.10.10"
+orjson==3.10.3 ; python_full_version == "3.10.10"
+overrides==7.7.0 ; python_full_version == "3.10.10"
+packaging==23.2 ; python_full_version == "3.10.10"
+pandas==2.2.2 ; python_full_version == "3.10.10"
+parso==0.8.4 ; python_full_version == "3.10.10"
+pexpect==4.9.0 ; sys_platform != "win32" and sys_platform != "emscripten" and python_full_version == "3.10.10"
+pillow==10.3.0 ; python_full_version == "3.10.10"
+platformdirs==4.2.1 ; python_full_version == "3.10.10"
+posthog==3.5.0 ; python_full_version == "3.10.10"
+prompt-toolkit==3.0.43 ; python_full_version == "3.10.10"
+protobuf==4.25.3 ; python_full_version == "3.10.10"
+psutil==5.9.8 ; python_full_version == "3.10.10"
+ptyprocess==0.7.0 ; sys_platform != "win32" and sys_platform != "emscripten" and python_full_version == "3.10.10"
+pulsar-client==3.5.0 ; python_full_version == "3.10.10"
+pure-eval==0.2.2 ; python_full_version == "3.10.10"
+pyasn1-modules==0.4.0 ; python_full_version == "3.10.10"
+pyasn1==0.6.0 ; python_full_version == "3.10.10"
+pycparser==2.22 ; implementation_name == "pypy" and python_full_version == "3.10.10"
+pydantic-core==2.18.2 ; python_full_version == "3.10.10"
+pydantic==2.7.1 ; python_full_version == "3.10.10"
+pydub==0.25.1 ; python_full_version == "3.10.10"
+pygments==2.18.0 ; python_full_version == "3.10.10"
+pyparsing==3.1.2 ; python_full_version == "3.10.10"
+pypika==0.48.9 ; python_full_version == "3.10.10"
+pyproject-hooks==1.1.0 ; python_full_version == "3.10.10"
+pyreadline3==3.4.1 ; sys_platform == "win32" and python_full_version == "3.10.10"
+pysqlite3-binary==0.5.2.post3 ; python_full_version == "3.10.10"
+python-dateutil==2.9.0.post0 ; python_full_version == "3.10.10"
+python-dotenv==1.0.1 ; python_full_version == "3.10.10"
+python-multipart==0.0.9 ; python_full_version == "3.10.10"
+pytz==2024.1 ; python_full_version == "3.10.10"
+pywin32==306 ; sys_platform == "win32" and platform_python_implementation != "PyPy" and python_full_version == "3.10.10"
+pyyaml==6.0.1 ; python_full_version == "3.10.10"
+pyzmq==26.0.3 ; python_full_version == "3.10.10"
+referencing==0.35.1 ; python_full_version == "3.10.10"
+regex==2024.5.10 ; python_full_version == "3.10.10"
+requests-file==2.0.0 ; python_full_version == "3.10.10"
+requests-oauthlib==1.3.1 ; python_full_version == "3.10.10"
+requests==2.31.0 ; python_full_version == "3.10.10"
+rich==13.7.1 ; python_full_version == "3.10.10"
+rpds-py==0.18.1 ; python_full_version == "3.10.10"
+rsa==4.9 ; python_full_version == "3.10.10"
+ruff==0.4.4 ; sys_platform != "emscripten" and python_full_version == "3.10.10"
+semantic-version==2.10.0 ; python_full_version == "3.10.10"
+setuptools==69.5.1 ; python_full_version == "3.10.10"
+sgmllib3k==1.0.0 ; python_full_version == "3.10.10"
+shellingham==1.5.4 ; python_full_version == "3.10.10"
+six==1.16.0 ; python_full_version == "3.10.10"
+sniffio==1.3.1 ; python_full_version == "3.10.10"
+soupsieve==2.5 ; python_full_version == "3.10.10"
+sqlalchemy==2.0.30 ; python_full_version == "3.10.10"
+stack-data==0.6.3 ; python_full_version == "3.10.10"
+starlette==0.37.2 ; python_full_version == "3.10.10"
+sympy==1.12 ; python_full_version == "3.10.10"
+tenacity==8.3.0 ; python_full_version == "3.10.10"
+tiktoken==0.6.0 ; python_full_version == "3.10.10"
+tinysegmenter==0.3 ; python_full_version == "3.10.10"
+tldextract==5.1.2 ; python_full_version == "3.10.10"
+tokenizers==0.19.1 ; python_full_version == "3.10.10"
+tomli==2.0.1 ; python_full_version == "3.10.10"
+tomlkit==0.12.0 ; python_full_version == "3.10.10"
+toolz==0.12.1 ; python_full_version == "3.10.10"
+tornado==6.4 ; python_full_version == "3.10.10"
+tqdm==4.66.4 ; python_full_version == "3.10.10"
+traitlets==5.14.3 ; python_full_version == "3.10.10"
+tweepy==4.14.0 ; python_full_version == "3.10.10"
+typer==0.12.3 ; python_full_version == "3.10.10"
+types-requests==2.31.0.20240406 ; python_full_version == "3.10.10"
+typing-extensions==4.11.0 ; python_full_version == "3.10.10"
+typing-inspect==0.9.0 ; python_full_version == "3.10.10"
+tzdata==2024.1 ; python_full_version == "3.10.10"
+ujson==5.9.0 ; python_full_version == "3.10.10"
+urllib3==2.2.1 ; python_full_version == "3.10.10"
+uvicorn==0.29.0 ; sys_platform != "emscripten" and python_full_version == "3.10.10"
+uvicorn[standard]==0.29.0 ; python_full_version == "3.10.10"
+uvloop==0.19.0 ; sys_platform != "win32" and sys_platform != "cygwin" and platform_python_implementation != "PyPy" and python_full_version == "3.10.10"
+watchfiles==0.21.0 ; python_full_version == "3.10.10"
+wcwidth==0.2.13 ; python_full_version == "3.10.10"
+websocket-client==1.8.0 ; python_full_version == "3.10.10"
+websockets==11.0.3 ; python_full_version == "3.10.10"
+wrapt==1.16.0 ; python_full_version == "3.10.10"
+yarl==1.9.4 ; python_full_version == "3.10.10"
+zipp==3.18.1 ; python_full_version == "3.10.10"