init w/ file upload

.gitignore

@@ -0,0 +1,5 @@
+.vscode
+__pycache__
+
+docker-compose.yml
+Dockerfile

app.py

@@ -0,0 +1,97 @@
+import os
+import gradio as gr
+
+from langchain.embeddings import HuggingFaceInstructEmbeddings
+from langchain.text_splitter import CharacterTextSplitter
+from langchain.vectorstores import Chroma
+from langchain.document_loaders import TextLoader
+
+
+def load_embeddings():
+    print(os.environ)
+    model_name = os.environ['HUGGINGFACEHUB_EMBEDDINGS_MODEL_NAME']
+    return HuggingFaceInstructEmbeddings(model_name=model_name)
+
+
+def split_file(file):
+    print(file.name)
+    loader = TextLoader(file.name)
+    documents = loader.load()
+    text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=20)
+    return text_splitter.split_documents(documents)
+
+
+def get_persist_directory(file_name):
+    return os.path.join(os.environ['CHROMADB_PERSIST_DIRECTORY'], file_name)
+
+
+def process_file(file):
+    embeddings = load_embeddings()
+    print(embeddings)
+    docs = split_file(file)
+    print(docs)
+
+    file_name, _ = os.path.splitext(os.path.basename(file.name))
+    persist_directory = get_persist_directory(file_name)
+    print(persist_directory)
+    vectordb = Chroma.from_documents(documents=docs, embedding=embeddings,
+                                     collection_name=file_name, persist_directory=persist_directory)
+    print(vectordb._client.list_collections())
+    vectordb.persist()
+    return None
+
+
+def load_vectordb(file_name):
+    embeddings = load_embeddings()
+
+    persist_directory = get_persist_directory(file_name)
+    vectordb = Chroma(collection_name=file_name,
+                      embedding_function=embeddings, persist_directory=persist_directory)
+    return vectordb
+
+
+def add_text(bot_history, text):
+    bot_history = bot_history + [(text, None)]
+    return bot_history, ""
+
+
+def bot(bot_history):
+    bot_history[-1][1] = 'so cool!'
+    return bot_history
+
+
+def clear_bot():
+    return None
+
+
+title = "QnA Chatbot"
+
+with gr.Blocks() as demo:
+    gr.Markdown(f"# {title}")
+
+    with gr.Row():
+        with gr.Column(scale=0.5):
+            upload = gr.File(file_types=["text"], label="Upload file")
+
+            process = gr.Button("Process")
+
+        with gr.Column(scale=0.5):
+            chatbot = gr.Chatbot([], elem_id="chatbot").style(height=750)
+
+            txt = gr.Textbox(
+                show_label=False,
+                placeholder="Enter text and press enter",
+            ).style(container=False)
+
+            clear = gr.Button("Clear")
+
+    process.click(process_file, upload, None)
+
+    txt.submit(add_text, [chatbot, txt], [chatbot, txt]).then(
+        bot, chatbot, chatbot
+    )
+    clear.click(clear_bot, None, chatbot)
+
+demo.title = title
+
+demo.launch()

requirements.txt

@@ -0,0 +1,5 @@
+langchain
+huggingface_hub
+sentence_transformers
+InstructorEmbedding
+chromadb