Conversation
Token-based scanner with English and Persian keywords. Extracts packages, imports, functions, structs/methods, enums, interfaces, impl, extern, layout and components, with call and type-reference edges and a cross-file resolver for package calls, receiver types and @fa/@en aliases.
Drop removed English/Persian spellings (while, old Persian words), add واردسازی, برابر/نابرابر, the repeat-step 'each'/هر rule, and Persian comma and question mark.
Match the current compiler: add switch/ترابرد statement parsing (multi-value,
range and relational cases, else), recognize «» guillemet string literals,
and stop a keyword reused as a struct-literal field name (Type { if = true })
from being misread as a control block opener.
|
Thanks for the pull request, @BaseMax. A maintainer will review it soon. Want to talk it through while it is in review? Come join us on our Discord server. For longer-form discussion there is also GitHub Discussions. A couple of things that speed up review: make sure the test suite passes on Python 3.10 and 3.13, and that the change keeps extraction deterministic. |
![graphify-labs[bot]](https://avatars.githubusercontent.com/in/4356270?s=60&v=4){kind=small}
There was a problem hiding this comment.
Graphify reviewed this change.
Worth a look — the grounded gate found no coupling regressions or blocking issues, but 5 advisory finding(s) below merit a look before merge.
Graphify review — findings
Adds a Salam (.salam) language extractor implemented as a standalone token scanner with no third-party parser, wiring it into detection, analysis family mapping, the extractor registry, dispatch, and the reference resolver via extract_salam and resolve_salam_references. The scanner handles English and Persian keyword spellings, switch/case, guillemet «» strings, and folds Arabic yeh/kaf and ZWNJ so Persian and English identifiers canonicalize, emitting packages, imports, functions, structs/methods, enums, interfaces, impl, extern C, layout/components, plus call and type-reference edges with @en/@fa name aliases so cross-language call sites resolve to their definitions. Documents the new language in the README support table.
Worth a look
- handle_impl returns None despite int cursor contract —
graphify/extractors/salam.py· Escalate · high- agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review
- Resolver raises on null arity metadata —
graphify/extractors/salam.py· Escalate · medium- agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review
- Variadic marker is counted as an extra parameter —
graphify/extractors/salam.py· Escalate · medium- agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review
- Trailing comma counted as an extra call argument —
graphify/extractors/salam.py· Escalate · medium- agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review
- Top-level variable type references are parsed then discarded —
graphify/extractors/salam.py· Escalate · medium- agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review
Analysis details — impact, health, verification
Impact & health
Graphify review
Impact — 3211 functions depend on the 555 functions this change touches.
Health — this change adds coupling hotspots:
- new:
extract()— 715 callers, 48 callees - new:
_rebuild_code()— 144 callers, 55 callees - new:
detect()— 112 callers, 15 callees - new:
to_obsidian()— 41 callers, 14 callees - new:
_extract_generic()— 18 callers, 29 callees - new:
save_manifest()— 42 callers, 12 callees - new:
to_json()— 58 callers, 7 callees - new:
extract_js()— 87 callers, 4 callees - …and 81 more — each is listed as a finding
Verification — 3211 functions in the blast radius were not formally verified this run (proofs are advisory here).
Gate & verification
graphify gate
PASS — objectively clean (no health regressions, tests not run — proofs not run this pass (advisory)). Grounded, not self-assessed.
Advisory (not blocking):
- verification_scope: 2891 function(s) in the blast radius were not formally verified this run
Test selection
Test selection
303 of 303 test file(s) selected (100%) via static blast radius.
Escalated to a full run for safety — the selection is not trustworthy on its own (see below). CI should run the whole suite.
tests/test_affected_cli.py— full-run-safetytests/test_affected_member_seed.py— full-run-safetytests/test_agents_platform.py— full-run-safetytests/test_analyze.py— impact, full-run-safetytests/test_anthropic_custom_endpoint.py— full-run-safetytests/test_antigravity_install.py— full-run-safetytests/test_apm_fallback_version.py— full-run-safetytests/test_architecture_doc.py— full-run-safetytests/test_astro_extraction.py— impact, full-run-safetytests/test_astro_import_ids.py— impact, full-run-safetytests/test_atomic_canvas_export.py— impact, full-run-safetytests/test_atomic_version_stamp.py— full-run-safetytests/test_atomic_writes.py— impact, full-run-safetytests/test_backend_env_isolation.py— full-run-safetytests/test_backend_extras.py— full-run-safetytests/test_benchmark.py— full-run-safetytests/test_benchmark_raw_graph.py— full-run-safetytests/test_build.py— impact, full-run-safetytests/test_build_merge_dedup_scope.py— full-run-safetytests/test_build_merge_hyperedges_and_prune.py— full-run-safetytests/test_build_merge_shrink_guard.py— full-run-safetytests/test_builtin_global_type_refs.py— impact, full-run-safetytests/test_cache.py— full-run-safetytests/test_callflow_html.py— full-run-safetytests/test_cargo_introspect.py— impact, full-run-safetytests/test_cargo_missing_manifest.py— full-run-safetytests/test_carried_hyperedge_remap.py— impact, full-run-safetytests/test_case_sensitive_resolution.py— impact, full-run-safetytests/test_charmap_encoding.py— impact, full-run-safetytests/test_chunking.py— impact, full-run-safetytests/test_cjs_module_extension.py— impact, full-run-safetytests/test_claude_cli_backend.py— impact, full-run-safetytests/test_claude_md.py— full-run-safetytests/test_cli_broken_pipe.py— full-run-safetytests/test_cli_export.py— impact, full-run-safetytests/test_cli_help.py— full-run-safetytests/test_cluster.py— full-run-safetytests/test_cobol_extractor.py— impact, full-run-safetytests/test_codebuddy.py— full-run-safetytests/test_community_hub_labels.py— full-run-safetytests/test_community_labels_skill.py— impact, full-run-safetytests/test_confidence.py— impact, full-run-safetytests/test_corrupt_graph_json.py— full-run-safetytests/test_cpp_nested_and_cli.py— impact, full-run-safetytests/test_cpp_objc_cross_file_calls.py— impact, full-run-safetytests/test_cpp_preprocess.py— full-run-safetytests/test_cross_extension_reexport_self_cycle.py— impact, full-run-safetytests/test_cross_language_call_resolution.py— impact, full-run-safetytests/test_cross_repo_external_call_guards.py— impact, full-run-safetytests/test_cross_repo_member_calls.py— impact, full-run-safety- … and 253 more
non-code file(s) changed (
README.md,tests/fixtures/sample.salam,tests/fixtures/sample_fa.salam) → running the full suite for safety (a code graph can't see config/fixture/data deps)
changed code file(s) with no mapped test (
README.md,graphify/extractors/__init__.py) — a coverage gap or a missing link — running the full suite rather than only the selected tests
Selection is safe under the controlled-regression assumption; always-run tests + a periodic full run are the backstops. Advisory — it never changes the check verdict.
· 89 more finding(s) on lines outside this diff (see the check run).
Handles makeCircle(r).area(), a local bound from a call, a struct literal, and an 'as Type' cast, using the callee's declared return type across files where needed. Deeper (2+ hop) chains stay unresolved rather than guessed.
# Conflicts: # CHANGELOG.md
There was a problem hiding this comment.
Graphify reviewed this change.
Worth a look — the grounded gate found no coupling regressions or blocking issues, but 5 advisory finding(s) below merit a look before merge.
Graphify review — findings
Adds a Salam (.salam) language extractor via extract_salam, a dependency-free token scanner that emits packages, imports, functions, structs with methods, enums, interfaces, impl blocks, type aliases, constants, extern: C functions, link libraries, and layout blocks/components, along with calls and references edges, and registers it across detection, dispatch, and the analyze language-family map. Wires up resolve_salam_references as a cross-file resolver that binds pkg.Func(), string-import aliases, receiver-typed method calls, and type references — reading a callee's declared return type across files to resolve member calls chained off another call, a local variable, a struct literal, or an as Type cast, while leaving 2+-hop chains unresolved rather than guessing. Handles both English and Persian keyword spellings (including switch/case, the «» guillemet string form, and the repeat ... each N in i step spelling) and avoids misreading a keyword reused as a struct-literal field name as a control block.
Worth a look
- Arity filter falls back to wrong-arity candidates —
graphify/extractors/salam.py· Escalate · medium- agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review
- Unqualified type refs resolve to unique type in unrelated package —
graphify/extractors/salam.py· Escalate · medium- agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review
- Qualified interface name is discarded in impl declarations —
graphify/extractors/salam.py· Escalate · medium- agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review
- Builtin casts overwrite earlier user-type casts —
graphify/extractors/salam.py· Escalate · medium- agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review
- Multi-token type names are parsed as only the first token —
graphify/extractors/salam.py· Escalate · medium- agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review
Analysis details — impact, health, verification
Impact & health
Graphify review
Impact — 3446 functions depend on the 790 functions this change touches.
Health — this change adds coupling hotspots:
- new:
extract()— 721 callers, 48 callees - new:
_rebuild_code()— 144 callers, 55 callees - new:
detect()— 112 callers, 15 callees - new:
to_obsidian()— 41 callers, 14 callees - new:
_extract_generic()— 18 callers, 29 callees - new:
save_manifest()— 42 callers, 12 callees - new:
to_json()— 58 callers, 7 callees - new:
extract_js()— 87 callers, 4 callees - …and 81 more — each is listed as a finding
Verification — 3446 functions in the blast radius were not formally verified this run (proofs are advisory here).
Gate & verification
graphify gate
PASS — objectively clean (no health regressions, tests not run — proofs not run this pass (advisory)). Grounded, not self-assessed.
Advisory (not blocking):
- verification_scope: 3126 function(s) in the blast radius were not formally verified this run
Test selection
Test selection
303 of 303 test file(s) selected (100%) via static blast radius.
Escalated to a full run for safety — the selection is not trustworthy on its own (see below). CI should run the whole suite.
tests/test_affected_cli.py— full-run-safetytests/test_affected_member_seed.py— full-run-safetytests/test_agents_platform.py— full-run-safetytests/test_analyze.py— impact, full-run-safetytests/test_anthropic_custom_endpoint.py— full-run-safetytests/test_antigravity_install.py— full-run-safetytests/test_apm_fallback_version.py— full-run-safetytests/test_architecture_doc.py— full-run-safetytests/test_astro_extraction.py— impact, full-run-safetytests/test_astro_import_ids.py— impact, full-run-safetytests/test_atomic_canvas_export.py— impact, full-run-safetytests/test_atomic_version_stamp.py— full-run-safetytests/test_atomic_writes.py— impact, full-run-safetytests/test_backend_env_isolation.py— full-run-safetytests/test_backend_extras.py— full-run-safetytests/test_benchmark.py— full-run-safetytests/test_benchmark_raw_graph.py— full-run-safetytests/test_build.py— impact, full-run-safetytests/test_build_merge_dedup_scope.py— full-run-safetytests/test_build_merge_hyperedges_and_prune.py— full-run-safetytests/test_build_merge_shrink_guard.py— full-run-safetytests/test_builtin_global_type_refs.py— impact, full-run-safetytests/test_cache.py— full-run-safetytests/test_callflow_html.py— full-run-safetytests/test_cargo_introspect.py— impact, full-run-safetytests/test_cargo_missing_manifest.py— full-run-safetytests/test_carried_hyperedge_remap.py— impact, full-run-safetytests/test_case_sensitive_resolution.py— impact, full-run-safetytests/test_charmap_encoding.py— impact, full-run-safetytests/test_chunking.py— impact, full-run-safetytests/test_cjs_module_extension.py— impact, full-run-safetytests/test_claude_cli_backend.py— impact, full-run-safetytests/test_claude_md.py— full-run-safetytests/test_cli_broken_pipe.py— full-run-safetytests/test_cli_export.py— impact, full-run-safetytests/test_cli_help.py— full-run-safetytests/test_cluster.py— full-run-safetytests/test_cobol_extractor.py— impact, full-run-safetytests/test_codebuddy.py— full-run-safetytests/test_community_hub_labels.py— full-run-safetytests/test_community_labels_skill.py— impact, full-run-safetytests/test_confidence.py— impact, full-run-safetytests/test_corrupt_graph_json.py— full-run-safetytests/test_cpp_nested_and_cli.py— impact, full-run-safetytests/test_cpp_objc_cross_file_calls.py— impact, full-run-safetytests/test_cpp_preprocess.py— full-run-safetytests/test_cross_extension_reexport_self_cycle.py— impact, full-run-safetytests/test_cross_language_call_resolution.py— impact, full-run-safetytests/test_cross_repo_external_call_guards.py— impact, full-run-safetytests/test_cross_repo_member_calls.py— impact, full-run-safety- … and 253 more
non-code file(s) changed (
CHANGELOG.md,README.md,tests/fixtures/sample.salam,tests/fixtures/sample_fa.salam) → running the full suite for safety (a code graph can't see config/fixture/data deps)
changed code file(s) with no mapped test (
CHANGELOG.md,README.md,graphify/extractors/__init__.py) — a coverage gap or a missing link — running the full suite rather than only the selected tests
Selection is safe under the controlled-regression assumption; always-run tests + a periodic full run are the backstops. Advisory — it never changes the check verdict.
· 89 more finding(s) on lines outside this diff (see the check run).
Bare function names passed as arguments (register("/", home)) now resolve
to indirect_call edges, same-file or package-qualified across files, with
shadowing by a same-named param/local correctly excluded.
Also: this/این was missing from the keyword table (receiver chains for
this.field could merge wrong), and a multi-word method name was truncated
to its last word on call.
There was a problem hiding this comment.
Graphify reviewed this change.
Worth a look — the grounded gate found no coupling regressions or blocking issues, but 2 advisory finding(s) below merit a look before merge.
Graphify review — findings
Adds a Salam (.salam) language extractor — a dependency-free token scanner (extract_salam) that emits packages, imports, functions, structs with methods, enums, interfaces, impl blocks, type aliases, constants, extern C functions, and layout/components across the full English and Persian keyword set (including switch/case, guillemet «» strings, and @en/@fa name aliases), wired into detection, dispatch, and the language-family maps. Its cross-file resolver (resolve_salam_references) binds pkg.Func() calls, string-import aliases, receiver-typed method calls, call-chain and cast return types, and struct-literal targets, resolving a bare function reference passed as an argument to an indirect_call edge (cross-file only when package-qualified, same-file otherwise to avoid wrong guesses). Fixes a param/local shadowing an import alias being misread as a package-qualified reference, restores multi-word method-name merging, and registers the omitted this/این keyword so this.field receiver chains resolve.
Worth a look
- Call resolution links functions with too many required parameters —
graphify/extractors/salam.py· Escalate · medium- agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review
- Public component flag is ignored —
graphify/extractors/salam.py· Escalate · medium- agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review
Analysis details — impact, health, verification
Impact & health
Graphify review
Impact — 3464 functions depend on the 808 functions this change touches.
Health — this change adds coupling hotspots:
- new:
extract()— 723 callers, 48 callees - new:
_rebuild_code()— 144 callers, 55 callees - new:
detect()— 112 callers, 15 callees - new:
to_obsidian()— 41 callers, 14 callees - new:
_extract_generic()— 18 callers, 29 callees - new:
save_manifest()— 42 callers, 12 callees - new:
to_json()— 58 callers, 7 callees - new:
extract_js()— 87 callers, 4 callees - …and 81 more — each is listed as a finding
Verification — 3464 functions in the blast radius were not formally verified this run (proofs are advisory here).
Gate & verification
graphify gate
PASS — objectively clean (no health regressions, tests not run — proofs not run this pass (advisory)). Grounded, not self-assessed.
Advisory (not blocking):
- verification_scope: 3144 function(s) in the blast radius were not formally verified this run
Test selection
Test selection
303 of 303 test file(s) selected (100%) via static blast radius.
Escalated to a full run for safety — the selection is not trustworthy on its own (see below). CI should run the whole suite.
tests/test_affected_cli.py— full-run-safetytests/test_affected_member_seed.py— full-run-safetytests/test_agents_platform.py— full-run-safetytests/test_analyze.py— impact, full-run-safetytests/test_anthropic_custom_endpoint.py— full-run-safetytests/test_antigravity_install.py— full-run-safetytests/test_apm_fallback_version.py— full-run-safetytests/test_architecture_doc.py— full-run-safetytests/test_astro_extraction.py— impact, full-run-safetytests/test_astro_import_ids.py— impact, full-run-safetytests/test_atomic_canvas_export.py— impact, full-run-safetytests/test_atomic_version_stamp.py— full-run-safetytests/test_atomic_writes.py— impact, full-run-safetytests/test_backend_env_isolation.py— full-run-safetytests/test_backend_extras.py— full-run-safetytests/test_benchmark.py— full-run-safetytests/test_benchmark_raw_graph.py— full-run-safetytests/test_build.py— impact, full-run-safetytests/test_build_merge_dedup_scope.py— full-run-safetytests/test_build_merge_hyperedges_and_prune.py— full-run-safetytests/test_build_merge_shrink_guard.py— full-run-safetytests/test_builtin_global_type_refs.py— impact, full-run-safetytests/test_cache.py— full-run-safetytests/test_callflow_html.py— full-run-safetytests/test_cargo_introspect.py— impact, full-run-safetytests/test_cargo_missing_manifest.py— full-run-safetytests/test_carried_hyperedge_remap.py— impact, full-run-safetytests/test_case_sensitive_resolution.py— impact, full-run-safetytests/test_charmap_encoding.py— impact, full-run-safetytests/test_chunking.py— impact, full-run-safetytests/test_cjs_module_extension.py— impact, full-run-safetytests/test_claude_cli_backend.py— impact, full-run-safetytests/test_claude_md.py— full-run-safetytests/test_cli_broken_pipe.py— full-run-safetytests/test_cli_export.py— impact, full-run-safetytests/test_cli_help.py— full-run-safetytests/test_cluster.py— full-run-safetytests/test_cobol_extractor.py— impact, full-run-safetytests/test_codebuddy.py— full-run-safetytests/test_community_hub_labels.py— full-run-safetytests/test_community_labels_skill.py— impact, full-run-safetytests/test_confidence.py— impact, full-run-safetytests/test_corrupt_graph_json.py— full-run-safetytests/test_cpp_nested_and_cli.py— impact, full-run-safetytests/test_cpp_objc_cross_file_calls.py— impact, full-run-safetytests/test_cpp_preprocess.py— full-run-safetytests/test_cross_extension_reexport_self_cycle.py— impact, full-run-safetytests/test_cross_language_call_resolution.py— impact, full-run-safetytests/test_cross_repo_external_call_guards.py— impact, full-run-safetytests/test_cross_repo_member_calls.py— impact, full-run-safety- … and 253 more
non-code file(s) changed (
CHANGELOG.md,README.md,tests/fixtures/sample.salam,tests/fixtures/sample_fa.salam) → running the full suite for safety (a code graph can't see config/fixture/data deps)
changed code file(s) with no mapped test (
CHANGELOG.md,README.md,graphify/extractors/__init__.py) — a coverage gap or a missing link — running the full suite rather than only the selected tests
Selection is safe under the controlled-regression assumption; always-run tests + a periodic full run are the backstops. Advisory — it never changes the check verdict.
· 89 more finding(s) on lines outside this diff (see the check run).
This branch has not been deployed
1 participant
What does this PR do?
Adds an extractor for Salam (
.salam) source files, so a Salam codebase gets extracted into the graph instead of being skipped. Salam has no tree-sitter grammar, so this is a small token-based scanner instead (like the COBOL/Apex extractors), no new dependency.Salam can be written in English or Persian (every keyword has both spellings), so the scanner recognizes both. It pulls out packages, imports, functions, structs + methods, enums, interfaces,
impl, type aliases, constants,externC functions,linklibraries, layout blocks/components, andswitch/match.@en/@faannotations become name aliases so a Persian call resolves to its English definition and back. Cross-file resolution covers package-qualified calls, string-import aliases, and receiver-typed method calls.Closes #3860
Type of change
Verification & Invariants
How was this tested?
Ran the extractor over the ~2600
.salamfiles in the Salam repo itself (English and Persian sources, stdlib and compiler) and checked every block opens/closes correctly. Addedtests/test_salam_extractor.py(16 tests) covering declarations, cross-file calls, switch/match, and Persian keywords. Ran the full test suite.Limitations: a chained call on an unresolved return type (f(x).method()) doesn't resolve, since the extractor doesn't track return types through call chains.
Graphify-specific checklist
[x] I updated generated skill artifacts (uv run python -m tools.skillgen --bless) when changing their source fragments.
uv run python -m tools.skillgen --check # 134 artifacts, no drift[x] I confirmed that AST/structural extraction remains deterministic (no ambient state dependencies like ENV variables).
[x] I reviewed changes for security implications (no unsafe interpolation into shell/Python).
[x] I confirmed no API keys or local-only graph data are included.
[x] I disclosed AI authorship in my commit messages.