Merge pull request #1 from NTTLuke/experiments

Experiments

.gitignore

@@ -158,9 +158,11 @@ cython_debug/
 #  and can be added to the global gitignore or merged into this file.  For a more nuclear
 #  option (not recommended) you can uncomment the following to ignore the entire idea folder.
 #.idea/
-chroma_db/chroma.sqlite3
-data/chroma/chroma.sqlite3
-data/chroma/863ec92d-3170-4dcc-b3ac-7b99a5b05b5e/data_level0.bin
-data/chroma/863ec92d-3170-4dcc-b3ac-7b99a5b05b5e/header.bin
-data/chroma/863ec92d-3170-4dcc-b3ac-7b99a5b05b5e/length.bin
-data/chroma/863ec92d-3170-4dcc-b3ac-7b99a5b05b5e/link_lists.bin
+chroma_db/*
+data/*
+.env_example
+docs/Perizia.pdf
+docs/detrazioni.txt
+sample-docs/detrazioni-miste.txt
+
+

.vscode/launch.json

@@ -10,7 +10,7 @@
       "request": "launch",
       "program": "${file}",
       "console": "integratedTerminal",
-      "justMyCode": true
+      "justMyCode": false
     }
   ]
 }

Pipfile

@@ -4,12 +4,27 @@ verify_ssl = true
 name = "pypi"
 
 [packages]
-langchain = "*"
 python-dotenv = "*"
 langsmith = "*"
 openai = "*"
-chromadb = "*"
 pypdf = "*"
+duckduckgo-search = "*"
+requests = "*"
+beautifulsoup4 = "*"
+langserve = "*"
+sse-starlette = "*"
+langchain = "*"
+docarray = "==0.39.1"
+pydantic = "==1.10.13"
+azure-search-documents = "*"
+azure-identity = "==1.15.0"
+chromadb = "*"
+llama-index = "*"
+traceloop-sdk = "*"
+langchain-openai = "*"
+sentence-transformers = "*"
+ragas = "*"
+langchainhub = "*"
 
 [dev-packages]
 ipykernel = "*"

analysis/main_bing_search.py

@@ -0,0 +1,160 @@
+from langchain.chat_models.azure_openai import AzureChatOpenAI
+from langchain.prompts import ChatPromptTemplate
+from langchain.schema.output_parser import StrOutputParser
+import requests
+from bs4 import BeautifulSoup
+from langchain.schema.runnable import RunnablePassthrough, RunnableLambda
+import os
+from dotenv import load_dotenv
+from langchain.utilities import BingSearchAPIWrapper
+import json
+
+load_dotenv()
+
+
+RESULT_PER_QUESTION = 3
+
+bing_search = BingSearchAPIWrapper()
+
+
+def web_search(query: str, num_results: int = RESULT_PER_QUESTION):
+    results = bing_search.results(query, num_results)
+    return [r["link"] for r in results]
+
+
+SUMMARY_TEMPLATE = """{text}
+-----------
+Using above text, answer in short the following question:
+
+> {question}
+-----------
+
+if the question cannot be answered, imply summarize the text. Include all factual information, numbers, statistics, etc.
+"""
+
+SUMMARY_PROMPT = ChatPromptTemplate.from_template(SUMMARY_TEMPLATE)
+
+
+llm = AzureChatOpenAI(
+    azure_deployment=os.getenv("OPENAI_CHAT_DEPLOYMENT_NAME"),
+    openai_api_type="azure",
+    temperature=0.0,
+)
+
+
+def scrape_text(url: str):
+    try:
+        response = requests.get(url)
+        if response.status_code == 200:
+            soup = BeautifulSoup(response.text, "html.parser")
+            text = soup.get_text(separator="", strip=True)
+            return text
+        else:
+            return f"Failed to scrape text : {response.status_code}"
+
+    except Exception as e:
+        print(e)
+        return f"Failed to scrape text {e}"
+
+
+# get the content of the page and use it as the context
+# url = "https://blog.langchain.dev/announcing-langsmith/"
+scrape_and_summarize_chain = (
+    RunnablePassthrough.assign(text=lambda x: scrape_text(x["url"])[:10000])
+    | SUMMARY_PROMPT
+    | llm
+    | StrOutputParser()
+)
+
+web_search_chain = (
+    RunnablePassthrough.assign(urls=lambda x: web_search(x["question"]))
+    | (lambda x: [{"question": x["question"], "url": u} for u in x["urls"]])
+    | scrape_and_summarize_chain.map()
+)
+
+
+SEARCH_PROMPT = ChatPromptTemplate.from_messages(
+    [
+        (
+            "user",
+            "Write 3 google search queries to search online that form an"
+            "objective opinion from the following: {question}\n",
+        ),
+        "You must respond with a list of strings in the following format:"
+        '["query1", "query2", "query3"]',
+    ]
+)
+
+search_question_chain = SEARCH_PROMPT | llm | StrOutputParser() | json.loads
+
+full_research_chain = (
+    search_question_chain
+    | (lambda x: [{"question": q} for q in x])
+    | web_search_chain.map()
+)
+
+
+WRITER_SYSTEM_PROMPT = "You are an AI critical thinker research assistant. Your sole purpose is to write well written, critically acclaimed, objective and structured reports on given text."
+
+# Report prompts from https://github.com/assafelovic/gpt-researcher/blob/master/gpt_researcher/master/prompts.py
+RESEARCH_REPORT_TEMPLATE = """Information:
+--------
+{research_summary}
+--------
+Using the above information, answer the following question or topic: "{question}" in a detailed report -- \
+The report should focus on the answer to the question, should be well structured, informative, \
+in depth, with facts and numbers if available and a minimum of 1,200 words.
+You should strive to write the report as long as you can using all relevant and necessary information provided.
+You must write the report with markdown syntax.
+You MUST determine your own concrete and valid opinion based on the given information. Do NOT deter to general and meaningless conclusions.
+Write all used source urls at the end of the report, and make sure to not add duplicated sources, but only one reference for each.
+You must write the report in apa format.
+Please do your best, this is very important to my career."""  # noqa: E501
+
+prompt = ChatPromptTemplate.from_messages(
+    [
+        ("system", WRITER_SYSTEM_PROMPT),
+        ("user", RESEARCH_REPORT_TEMPLATE),
+    ]
+)
+
+
+def collapse_list_of_lists(list_of_lists):
+    content = []
+    for l in list_of_lists:
+        content.append("\n\n".join(l))
+    return "\n\n".join(content)
+
+
+chain = (
+    RunnablePassthrough.assign(
+        research_summary=full_research_chain | collapse_list_of_lists
+    )
+    | prompt
+    | llm
+    | StrOutputParser()
+)
+
+#####  LANGSERVE  #####
+#!/usr/bin/env python
+from fastapi import FastAPI
+from langserve import add_routes
+
+
+app = FastAPI(
+    title="LangChain Server",
+    version="1.0",
+    description="A simple api server using Langchain's Runnable interfaces",
+)
+
+add_routes(
+    app,
+    chain,
+    path="/research-assistant",
+)
+
+
+if __name__ == "__main__":
+    import uvicorn
+
+    uvicorn.run(app, host="localhost", port=8000)