Sweep: create a new agent to be used in ticket_utils.py

[wwzeng1]

Details

The agent should filter unnecessary terms out of the search query to be sent into lexical search. Write a prompt to do this, using pr_description as a reference

Checklist

• Create sweepai/agents/query_filter_agent.py ✓ 9159660 Edit
• Running GitHub Actions for sweepai/agents/query_filter_agent.py ✓ Edit
• Modify sweepai/utils/ticket_utils.py ✓ 8d22520 Edit
• Running GitHub Actions for sweepai/utils/ticket_utils.py ✓ Edit

[sweep-nightly]

🚀 Here's the PR! #3160

See Sweep's progress at the progress dashboard!

💎 Sweep Pro: I'm using GPT-4. You have unlimited GPT-4 tickets. (tracking ID: None)

---

Tip

I can email you next time I complete a pull request if you set up your email here!

---

Actions (click)

• ↻ Restart Sweep

---

Step 1: 🔎 Searching

I found the following snippets in your repository. I will now analyze these snippets and come up with a plan.

Some code snippets I think are relevant in decreasing order of relevance (click to expand). If some file is missing from here, you can mention the path in the ticket description.

sweep/sweepai/utils/ticket_utils.py

Lines 93 to 130 in 6430b99

	def fetch_relevant_files(
	cloned_repo,
	title,
	summary,
	replies_text,
	username,
	metadata,
	on_ticket_start_time,
	tracking_id,
	is_paying_user,
	is_consumer_tier,
	issue_url,
	chat_logger,
	ticket_progress: TicketProgress,
	):
	logger.info("Fetching relevant files...")
	try:
	search_query = (title + summary + replies_text).strip("\n")
	replies_text = f"\n{replies_text}" if replies_text else ""
	formatted_query = (f"{title.strip()}\n{summary.strip()}" + replies_text).strip(
	"\n"
	)
	repo_context_manager = prep_snippets(cloned_repo, search_query, ticket_progress)
	ticket_progress.search_progress.repo_tree = str(repo_context_manager.dir_obj)
	ticket_progress.save()
	repo_context_manager = get_relevant_context(
	formatted_query,
	repo_context_manager,
	ticket_progress,
	chat_logger=chat_logger,
	)
	snippets = repo_context_manager.current_top_snippets
	ticket_progress.search_progress.repo_tree = str(repo_context_manager.dir_obj)
	ticket_progress.search_progress.final_snippets = snippets
	ticket_progress.save()

sweep/sweepai/agents/assistant_functions.py

Lines 52 to 69 in 6430b99

	keyword_search_schema = {
	"name": "keyword_search",
	"parameters": {
	"type": "object",
	"properties": {
	"justification": {
	"type": "string",
	"description": "Justification for searching the keyword.",
	},
	"keyword": {
	"type": "string",
	"description": "The keyword to search for.",
	},
	},
	"required": ["justification", "keyword"],
	},
	"description": "Searches for all lines in the file containing the keyword.",

sweep/sweepai/agents/pr_description_bot.py

Lines 25 to 51 in 6430b99

	class PRDescriptionBot(ChatGPT):
	def describe_diffs(
	self,
	diffs,
	pr_title,
	):
	self.messages = []
	self.model = DEFAULT_GPT35_MODEL
	# attempt to generate description 3 times
	pr_desc_pattern = r"<pr_description>\n(.*?)\n</pr_description>"
	for attempt in [0, 1, 2]:
	pr_desc_response = self.chat(
	content=prompt.format(
	diffs=diffs,
	pr_title=pr_title,
	),
	)
	pr_desc_matches = re.search(pr_desc_pattern, pr_desc_response, re.DOTALL)
	if pr_desc_matches is None:
	if attempt == 2:
	return ""
	else:
	break
	pr_desc = pr_desc_matches.group(1)
	pr_desc = pr_desc.strip()

---

Step 2: ⌨️ Coding

• Create sweepai/agents/query_filter_agent.py ✓ 9159660 Edit

Create sweepai/agents/query_filter_agent.py with contents:
• Create a new Python file named query_filter_agent.py in the sweepai/agents directory.
• Define a new class named QueryFilterAgent within this file.
• Import necessary modules and classes, such as the ChatGPT class from sweepai/agents/pr_description_bot.py, as this new agent will likely use similar functionality.
• Implement a method within QueryFilterAgent named filter_search_query that takes a search query as input and returns a filtered version of the query. This method will use a prompt to filter out unnecessary terms, similar to how PRDescriptionBot generates PR descriptions.
• The method should handle cases where the filtering does not yield a result by returning the original query as a fallback.

• Running GitHub Actions for sweepai/agents/query_filter_agent.py ✓ Edit

Check sweepai/agents/query_filter_agent.py with contents:

Ran GitHub Actions for 9159660fe39e8db1ec039e836362627e567045b8:

• Modify sweepai/utils/ticket_utils.py ✓ 8d22520 Edit

Modify sweepai/utils/ticket_utils.py with contents:
• Import the QueryFilterAgent class from the newly created file sweepai/agents/query_filter_agent.py at the top of the ticket_utils.py file.
• Inside the fetch_relevant_files function, replace the current search_query concatenation logic (lines 111-115) with a call to the filter_search_query method of a QueryFilterAgent instance.
• Create an instance of QueryFilterAgent before the try block where the search_query is currently being concatenated.
• Pass the concatenated title, summary, and replies_text to the filter_search_query method and assign the result back to the search_query variable.
• Ensure that the rest of the function uses this new filtered search_query for further processing.

--- 
+++ 
@@ -5,6 +5,7 @@
 
 from sweepai.config.client import SweepConfig
 from sweepai.core.context_pruning import RepoContextManager, get_relevant_context
+from sweepai.agents.query_filter_agent import QueryFilterAgent
 from sweepai.core.lexical_search import (
     compute_vector_search_scores,
     prepare_lexical_search_index,
@@ -108,7 +109,8 @@
 ):
     logger.info("Fetching relevant files...")
     try:
-        search_query = (title + summary + replies_text).strip("\n")
+        filter_agent = QueryFilterAgent()
+        search_query = filter_agent.filter_search_query((title + summary + replies_text).strip('\n'))
         replies_text = f"\n{replies_text}" if replies_text else ""
         formatted_query = (f"{title.strip()}\n{summary.strip()}" + replies_text).strip(
             "\n"

• Running GitHub Actions for sweepai/utils/ticket_utils.py ✓ Edit

Check sweepai/utils/ticket_utils.py with contents:

Ran GitHub Actions for 8d2252011a45e78dce97cddb27a991e025f7568c:

---

Step 3: 🔁 Code Review

I have finished reviewing the code for completeness. I did not find errors for sweep/create_a_new_agent_to_be_used_in_ticket_2f9c3.

---

🎉 Latest improvements to Sweep:

• New dashboard launched for real-time tracking of Sweep issues, covering all stages from search to coding.
• Integration of OpenAI's latest Assistant API for more efficient and reliable code planning and editing, improving speed by 3x.
• Use the GitHub issues extension for creating Sweep issues directly from your editor.

💡 To recreate the pull request edit the issue title or description.
^{Something wrong? Let us know.}

This is an automated message generated by Sweep AI.