Muhammad Faraz Malik
Add Streamlit demo UI with scripted agent; merge_attempts reducer; docs
eb60226 Download tests/unit/test_mock_graph.py from MF180/autonomous-bug-fixing-agent: direct link, hf CLI and curl.
- Browser
- Download file 4.78 kB
-
https://huggingface.co/MF180/autonomous-bug-fixing-agent/resolve/main/tests/unit/test_mock_graph.py
- Command line
-
hf download hf://MF180/autonomous-bug-fixing-agent/tests/unit/test_mock_graph.py
-
curl -L -o test_mock_graph.py https://huggingface.co/MF180/autonomous-bug-fixing-agent/resolve/main/tests/unit/test_mock_graph.py
4.78 kB
| """Tests for the scripted demo graph: control flow, HITL pause, and the attempts reducer.""" | |
| from __future__ import annotations | |
| from pathlib import Path | |
| import pytest | |
| from langgraph.types import Command | |
| from agent.state import ( | |
| AgentStatus, | |
| FixAttempt, | |
| IssueInput, | |
| SandboxResult, | |
| initial_state, | |
| merge_attempts, | |
| ) | |
| from ui.mock_graph import DEMO_ISSUE_BODY, DEMO_ISSUE_TITLE, build_mock_graph | |
| CFG = {"configurable": {"thread_id": "t", "delay": 0}} | |
| def graph(): | |
| return build_mock_graph() | |
| def start_state(): | |
| return initial_state( | |
| IssueInput(title=DEMO_ISSUE_TITLE, body=DEMO_ISSUE_BODY, repo_path=Path("demo")) | |
| ) | |
| def _nodes(graph, payload, cfg=CFG) -> list[str]: | |
| return [n for chunk in graph.stream(payload, cfg, stream_mode="updates") for n in chunk] | |
| def _run_to_approval(graph, start_state): | |
| nodes = _nodes(graph, start_state) | |
| return nodes, graph.get_state(CFG) | |
| # ----------------------------------------------------------------------- reducer | |
| def _attempt(i: int, rationale: str = "", result: SandboxResult | None = None) -> FixAttempt: | |
| return FixAttempt(iteration=i, rationale=rationale, result=result) | |
| def test_merge_attempts_appends_new_iterations_in_order(): | |
| merged = merge_attempts([_attempt(1)], [_attempt(2)]) | |
| assert [a.iteration for a in merged] == [1, 2] | |
| def test_merge_attempts_upserts_same_iteration_keeping_history(): | |
| res = SandboxResult(exit_code=1, stdout="F", stderr="") | |
| merged = merge_attempts([_attempt(1, "a"), _attempt(2, "b")], [_attempt(2, "b", res)]) | |
| assert [a.iteration for a in merged] == [1, 2] | |
| assert merged[1].result == res and merged[0].result is None | |
| def test_merge_attempts_handles_none(): | |
| assert merge_attempts(None, None) == [] | |
| # ------------------------------------------------------------------- control flow | |
| def test_graph_pauses_before_commit_and_a_failed_first_attempt_is_retried(graph, start_state): | |
| nodes, snap = _run_to_approval(graph, start_state) | |
| assert snap.next == ("human_approval",) # paused, waiting for a human | |
| assert "commit_and_pr" not in nodes | |
| assert nodes.count("propose_fix") == 2 and nodes.count("analyze_failure") == 1 | |
| assert snap.values["iteration"] == 2 | |
| attempts = snap.values["attempts"] | |
| assert [a.result.passed for a in attempts] == [False, True] # history kept, results attached | |
| assert snap.values.get("commit_sha") is None | |
| def test_interrupt_payload_contains_review_material(graph, start_state): | |
| _run_to_approval(graph, start_state) | |
| payload = graph.get_state(CFG).tasks[0].interrupts[0].value | |
| assert "+ return sum(xs) / len(xs)" in payload["diff"] | |
| assert payload["test_output"].startswith("2 passed") | |
| assert "def test_average_of_three" in payload["repro_test"] | |
| def test_approve_commits(graph, start_state): | |
| _run_to_approval(graph, start_state) | |
| nodes = _nodes(graph, Command(resume={"approved": True, "reviewer": "me"})) | |
| assert nodes[-1] == "commit_and_pr" | |
| final = graph.get_state(CFG) | |
| assert final.next == () | |
| assert final.values["status"] == AgentStatus.COMMITTED | |
| assert len(final.values["commit_sha"]) == 7 | |
| assert final.values["approval"].reviewer == "me" | |
| def test_request_changes_loops_back_then_pauses_again(graph, start_state): | |
| _run_to_approval(graph, start_state) | |
| nodes = _nodes(graph, Command(resume={"approved": False, "feedback": "handle empty list"})) | |
| assert "analyze_failure" in nodes and "propose_fix" in nodes | |
| assert "commit_and_pr" not in nodes | |
| snap = graph.get_state(CFG) | |
| assert snap.next == ("human_approval",) # a second approval is required | |
| assert snap.values["iteration"] == 3 | |
| assert "ValueError" in snap.values["attempts"][-1].patches[0].unified_diff | |
| def test_reject_without_feedback_escalates_and_never_commits(graph, start_state): | |
| _run_to_approval(graph, start_state) | |
| nodes = _nodes(graph, Command(resume={"approved": False})) | |
| assert nodes[-1] == "escalate" and "commit_and_pr" not in nodes | |
| final = graph.get_state(CFG).values | |
| assert final["status"] == AgentStatus.ESCALATED | |
| assert final["commit_sha"] is None and "rejected" in final["error"].lower() | |
| def test_repeated_change_requests_stop_at_iteration_budget(graph, start_state): | |
| _run_to_approval(graph, start_state) | |
| for _ in range(20): | |
| if not graph.get_state(CFG).next: | |
| break | |
| _nodes(graph, Command(resume={"approved": False, "feedback": "again"})) | |
| final = graph.get_state(CFG) | |
| assert final.next == () | |
| assert final.values["iteration"] <= final.values["budget"].max_iterations == 8 | |
| assert final.values["status"] == AgentStatus.ESCALATED | |
| assert final.values["commit_sha"] is None | |