[Hotfix] Address URL Formatting Issue Did Not Get Fixed During 0.3.0 (#…

…312)

### Description
- A previous PR didn't test thoroughly and had a bug so that it did not
address the problem.
- Previous PR: #306 
- Goal is to have different system prompt for different slack vs webapp
client
- The previous PR was modifying the user's question prompt not system
prompt

### Related Issue
closes #301 
This issue is supposed to be fixed in 0.3.0, but due to inadequate
testing during that PR merge, it wasn't caught that the PR didn't
address the problem. The bug is re-discovered after 0.3.0 release.

### Tests
1. Sent a request on front end
2. Got URL embedded into the keywords
<img width="1456" alt="image"
src="https://github.com/astronomer/ask-astro/assets/26350341/1f93cecf-6ff4-40db-90bd-6e055330e3ce">
3. Check Langsmith,  metadata has client = webapp
<img width="904" alt="image"
src="https://github.com/astronomer/ask-astro/assets/26350341/df4c2e26-ddeb-4b33-b1f7-f32f7009bafa">
4. Check langsmith, correct system prompt is used
<img width="1044" alt="image"
src="https://github.com/astronomer/ask-astro/assets/26350341/3245201d-73ed-4ebc-b4fe-7deff770cc4a">

[Langsmith
link](https://smith.langchain.com/public/37f44ef4-af11-45a5-8246-e7f2b1997b58/r)

airflow/include/streamlit/streamlit_app.py

@@ -138,7 +138,7 @@ def write_response(text: str):
              question to see the prompt."""
     )
 
-    prompt_path = Path(__file__).parent.joinpath("combine_docs_chat_prompt.txt")
+    prompt_path = Path(__file__).parent.joinpath("combine_docs_sys_prompt_webapp.txt")
     prompt = prompt_path.read_text()
 
     st.write(f"The prompt text can be found in the file __`{prompt_path}`__")

api/ask_astro/chains/answer_question.py

@@ -29,10 +29,18 @@
     MULTI_QUERY_RETRIEVER_TEMPERATURE,
 )
 
-with open("ask_astro/templates/combine_docs_chat_prompt.txt") as system_prompt_fd:
-    """Load system prompt template from a file and structure it."""
-    messages = [
-        SystemMessagePromptTemplate.from_template(system_prompt_fd.read()),
+with open("ask_astro/templates/combine_docs_sys_prompt_webapp.txt") as webapp_system_prompt_fd:
+    """Load system prompt template for webapp messages"""
+    webapp_messages = [
+        SystemMessagePromptTemplate.from_template(webapp_system_prompt_fd.read()),
+        MessagesPlaceholder(variable_name="messages"),
+        HumanMessagePromptTemplate.from_template("{question}"),
+    ]
+
+with open("ask_astro/templates/combine_docs_sys_prompt_slack.txt") as slack_system_prompt_fd:
+    """Load system prompt template for slack messages"""
+    slack_messages = [
+        SystemMessagePromptTemplate.from_template(slack_system_prompt_fd.read()),
         MessagesPlaceholder(variable_name="messages"),
         HumanMessagePromptTemplate.from_template("{question}"),
     ]
@@ -92,7 +100,29 @@
 )
 
 # Set up a ConversationalRetrievalChain to generate answers using the retriever.
-answer_question_chain = ConversationalRetrievalChain(
+webapp_answer_question_chain = ConversationalRetrievalChain(
+    retriever=llm_chain_filter_compression_retriever,
+    return_source_documents=True,
+    question_generator=LLMChain(
+        llm=AzureChatOpenAI(
+            **AzureOpenAIParams.us_east2,
+            deployment_name=CONVERSATIONAL_RETRIEVAL_LLM_CHAIN_DEPLOYMENT_NAME,
+            temperature=CONVERSATIONAL_RETRIEVAL_LLM_CHAIN_TEMPERATURE,
+        ),
+        prompt=CONDENSE_QUESTION_PROMPT,
+    ),
+    combine_docs_chain=load_qa_chain(
+        AzureChatOpenAI(
+            **AzureOpenAIParams.us_east2,
+            deployment_name=CONVERSATIONAL_RETRIEVAL_LOAD_QA_CHAIN_DEPLOYMENT_NAME,
+            temperature=CONVERSATIONAL_RETRIEVAL_LOAD_QA_CHAIN_TEMPERATURE,
+        ),
+        chain_type="stuff",
+        prompt=ChatPromptTemplate.from_messages(webapp_messages),
+    ),
+)
+
+slack_answer_question_chain = ConversationalRetrievalChain(
     retriever=llm_chain_filter_compression_retriever,
     return_source_documents=True,
     question_generator=LLMChain(
@@ -110,6 +140,6 @@
             temperature=CONVERSATIONAL_RETRIEVAL_LOAD_QA_CHAIN_TEMPERATURE,
         ),
         chain_type="stuff",
-        prompt=ChatPromptTemplate.from_messages(messages),
+        prompt=ChatPromptTemplate.from_messages(slack_messages),
     ),
 )

api/ask_astro/services/questions.py

@@ -82,7 +82,7 @@ async def answer_question(request: AskAstroRequest) -> None:
     try:
         from langchain import callbacks
 
-        from ask_astro.chains.answer_question import answer_question_chain
+        from ask_astro.chains.answer_question import slack_answer_question_chain, webapp_answer_question_chain
 
         # First, mark the request as in_progress and add it to the database
         request.status = "in_progress"
@@ -93,16 +93,27 @@ async def answer_question(request: AskAstroRequest) -> None:
 
         # Run the question answering chain
         with callbacks.collect_runs() as cb:
-            result = await asyncio.to_thread(
-                lambda: answer_question_chain(
-                    {
-                        "question": request.prompt,
-                        "chat_history": [],
-                        "messages": request.messages,
-                    },
-                    metadata={"request_id": str(request.uuid)},
+            if request.client == "slack":
+                result = await asyncio.to_thread(
+                    lambda: slack_answer_question_chain(
+                        {
+                            "question": request.prompt,
+                            "chat_history": request.messages,
+                        },
+                        metadata={"request_id": str(request.uuid), "client": str(request.client)},
+                    )
+                )
+            else:
+                result = await asyncio.to_thread(
+                    lambda: webapp_answer_question_chain(
+                        {
+                            "question": request.prompt,
+                            "chat_history": [],
+                            "messages": request.messages,
+                        },
+                        metadata={"request_id": str(request.uuid), "client": str(request.client)},
+                    )
                 )
-            )
             request.langchain_run_id = cb.traced_runs[0].id
 
         logger.info("Question answering chain finished with result %s", result)

api/ask_astro/templates/combine_docs_sys_prompt_webapp.txt

@@ -0,0 +1,17 @@
+You are Ask Astro, a friendy and helpful bot.
+Only answer questions related to Astronomer, the Astro platform and Apache Airflow.
+If the question relates to pricing, licensing, or commercial usage, ask the user to contact support at www.astronomer.io/contact.
+If you don't know the answer, just say that you don't know and ask the user to contact support, don't try to make up an answer.
+If the supplied context below does not have sufficient information to help answer the question, make a note when answering to let the user know that the answer may contain false information and the user should contact support to verify.
+Be concise and precise in your answers and do not apologize.
+Format your response using Markdown syntax.
+Surround text with SINGLE * to format it in bold or provide emphasis. Examples: GOOD: *This is bold!*. BAD: **This is bold!**.
+Support text with _ to format it in italic. Example: _This is italic._
+Use the • character for unnumbered lists.
+Use the ` character to surround inline code. Example: This is a sentence with some `inline *code*` in it.
+Use ``` to surround multi-line code blocks. Do not specify a language in code blocks. Examples: GOOD: ```This is a code block\nAnd it is multi-line``` BAD: ```python print("Hello world!")```.
+Format links using this format: [Text to display](URL). Examples: GOOD: [This message **is** a link](https://www.example.com). BAD: <https://www.example.com|This message **is** a link>.
+12 character words that start with "<@U" and end with ">" are usernames. Example: <@U024BE7LH>.
+Use the following pieces of context to answer the users question.
+----------------
+{context}

tests/api/ask_astro/chains/test_answer_questions.py

@@ -3,7 +3,7 @@
 
 def test_system_prompt_loading():
     """Test if the system prompt is loaded correctly"""
-    with open("ask_astro/templates/combine_docs_chat_prompt.txt") as fd:
+    with open("ask_astro/templates/combine_docs_sys_prompt_webapp.txt") as fd:
         expected_template = fd.read()
     template_instance = SystemMessagePromptTemplate.from_template(expected_template)
     assert template_instance.prompt.template == expected_template