* feat: detect circular import dependencies at file level
- Add find_import_cycles() to analyze.py
- Collapses symbol nodes to parent files, builds directed file graph
- Uses nx.simple_cycles() bounded by max_cycle_length (default 5)
- Deduplicates rotations, returns shortest cycles first
- Considers both imports_from and re_exports edges
Tested on a 976-file Next.js codebase: found 4 cycles including
a known utils↔barrel circular dependency and a 4-file API cycle.
* fix: resolve import-cycle merge blockers
- use source_file-only endpoint resolution (no label fallback)
- support Graph/DiGraph orientation via edge source_file
- return structured cycle records and include self-loops
- integrate Import Cycles section into GRAPH_REPORT.md
- expand cycle tests for real-schema IDs, undirected input,
missing source_file nodes, and non-import relations
Three compounding bugs caused ~30-50% of semantic chunks to come back
as 'hollow responses' on the claude-cli backend, triggering adaptive
bisection that doubled or tripled the number of subprocess calls.
Root causes
-----------
1. _parse_llm_json only stripped markdown fences when raw.startswith('```').
Claude frequently prepends a short preamble before the fence
('Here are the extracted entities:\n\n```json\n{...}```'), making
the check fail. json.loads then drops the chunk. Each bisected half
may exhibit the same failure, so cost compounds.
2. _call_claude_cli used --append-system-prompt, which layers graphify's
extraction prompt on top of Claude Code's default interactive-agent
prompt ('use markdown formatting', 'output text to communicate with
the user'). The conflicting instructions explain ~50% of the
preambles and fences from (1). Switching to --system-prompt (replace)
eliminates the conflict at the source.
3. claude-cli defaults to Opus, which is overkill for the structured
JSON extraction graphify performs. New GRAPHIFY_CLAUDE_CLI_MODEL env
var lets users opt into haiku / sonnet for big builds. Default
behaviour unchanged when the env var is unset.
Fix
---
- Robust _parse_llm_json: strips fences regardless of position, with a
balanced-brace fallback that scans for the first complete JSON object
in the response. Handles preambles, trailing prose, prose-wrapped
JSON without fences. Diagnostic log on terminal failure includes the
first 200 chars of the response.
- _call_claude_cli switches to --system-prompt.
- _call_claude_cli respects GRAPHIFY_CLAUDE_CLI_MODEL when set.
Tests (tests/test_llm_parser.py)
--------------------------------
- The four PR-body failure modes: preamble+fence, prose+JSON, raw JSON,
total refusal.
- Bonus: uppercase fence tag, unclosed fence, empty response.
- argv shape: --system-prompt present, --append-system-prompt absent.
- argv shape: --model added iff GRAPHIFY_CLAUDE_CLI_MODEL is set.
19/19 tests pass (9 pre-existing in test_claude_cli_backend.py +
10 new). Verified end-to-end on a 800-file repo: 0 hollow responses
after, vs ~30-50% before; output tokens -93%; wall time 44 min -> 4 min.
* chore: declare pytest as a uv dev dependency
The contributing guide currently tells contributors `pip install pytest`
as a separate step, and CI does the same. Move pytest into PEP 735
`[dependency-groups]` so it's declared in pyproject.toml and `uv sync`
installs it by default (no `--with` workaround, no separate install
line). Update CI to use astral-sh/setup-uv + `uv sync` + `uv run pytest`,
and refresh the Contributing section of the README to match.
`[dependency-groups]` is the right home (vs `[project.optional-dependencies]`)
because pytest is dev-only and shouldn't appear in the published wheel's
optional features list alongside things like `pdf` or `mcp`.
* remove uv.lock from gitignore
When a graph exceeds the viz node limit, to_html() builds a
community-aggregated meta-graph and recursively calls itself.
The recursive call never carried hyperedges onto the meta-graph,
so graph.html always emitted const hyperedges = [] even when
graph.json contained plenty.
This fix remaps hyperedge node references from semantic node IDs
to community IDs before the recursive call, so hyperedge regions
render correctly in the aggregated view. Hyperedges that collapse
to fewer than 2 distinct communities are dropped (they wouldn't
render as a polygon anyway).
Fixes#1005
graphify-out regenerates differently on every `graphify update` even
when no source changed, so the committed graph is perpetually dirty and
the post-commit/post-checkout hooks fight every commit. Two independent
nondeterminism sources, each fixed here:
1. Edge direction flips. build.py builds an undirected graph and stores
direction in _src/_tgt; collapsing two edges onto the same node pair
is last-write-wins, and unstable edge iteration order flips them
run-to-run. Fixed by sorting edges by (source, target, relation)
before the add loop.
2. Clustering churn. The networkx Louvain fallback iterates string-keyed
sets whose order is randomized per-process by PYTHONHASHSEED, so
community assignments differ run-to-run even with seed=42. Fixed by
exporting PYTHONHASHSEED=0 in the generated post-commit and
post-checkout hook scripts.
With both fixes, `graphify update` is idempotent: rebuilding an
already-converged graphify-out reproduces graph.json and GRAPH_REPORT.md
byte-for-byte.
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
stdlib ET does not cap entity expansion — a crafted .csproj or .lpk with nested
internal entities can exhaust memory. Pre-screen input bytes for <!DOCTYPE and
<!ENTITY before parsing (legitimate MSBuild/Lazarus files never contain these).
Also adds the missing 2 MiB size cap to extract_lpk (csproj already had one).
No new dependencies required.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds graphify/mcp_ingest.py — extracts MCP server configurations into the
knowledge graph. Captures server nodes, NuGet/npm/pip package refs, commands,
env var requirements, and inter-server edges. Dispatched by filename before
the suffix lookup so generic .json extraction is unaffected. Env values are
discarded to prevent secret leakage. File size capped at 1 MiB. 29 tests.
Fixes: server_count budget now checked after validity guard so invalid entries
don't consume capacity; removed misleading uv run docstring example.
Co-Authored-By: adityachaudhary99 <adityachaudhary99@users.noreply.github.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
cluster-only re-runs Leiden clustering and then re-applies the existing
.graphify_labels.json by raw cid index, which causes labels to attach to
clusters whose members are unrelated to the label's original meaning
whenever the graph has changed between labeling and re-clustering.
Mirror the safety net already present in watch.py:_rebuild_code added in
#822 for the watch/update paths.
Adds a regression test that fails without the fix (label cids become
orphaned from graph.json community attributes after re-clustering).
Refs: #1027
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Three issues in _rebuild_code (watch.py):
1. _relativize_source_files was called on result after eviction list was built,
so existing nodes with absolute source_file were never normalized before comparison
2. deleted_paths and evict_sources used str() (backslashes on Windows) while
graph.json stores forward-slash paths via _norm_source_file
3. _relativize_source_files itself used str() instead of as_posix()
Also fix extract.py source_file relativization to use as_posix(). Closes#1007.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace inlined path normalisation with _norm_source_file (the same function
that builds node source_file keys) so prune_set and node attrs are normalised
identically. resolve() on root handles symlinked scan roots. Keep both raw and
normalised forms in prune_set so nodes with absolute source_file also match.
Add edge pruning and Windows backslash path tests per Opus review.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
prune_set in build_merge now includes relative-path variants of each deleted file
so manifest absolute paths (e.g. /home/user/corpus/module_b/utils.py) match graph
node source_file values (e.g. module_b/utils.py) regardless of OS or run context.
Fixes#1007.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Keep Graphify query segmentation focused on Chinese terms: rename the CJK helpers and extra to Chinese scope, cache the optional jieba import at module load, and keep a bigram fallback when jieba is unavailable.
Constraint: Reviewer asked either to broaden Hiragana/Katakana/Hangul support or rename CJK helpers; user chose Chinese-only because Japanese segmentation accuracy is uncertain.
Rejected: Broaden to Japanese and Korean segmentation | jieba is Chinese-oriented and the user explicitly limited scope to Chinese.
Confidence: high
Scope-risk: narrow
Directive: Do not label this path as CJK unless Hiragana/Katakana/Hangul segmentation is intentionally supported and tested.
Tested: uv run --with pytest pytest tests/test_serve.py tests/test_query_cli.py tests/test_benchmark.py
Tested: uv run --with pytest --with jieba pytest tests/test_serve.py -k "chinese or non_chinese"
Tested: graphify update .
Not-tested: Full test suite.
Co-authored-by: OmX <omx@oh-my-codex.dev>
Safi