Is your feature request related to a problem? Please describe.
A new codeanalyzer-python release (v1.5.2) is published. Update the codeanalyzer-python pin in this repo and adapt the integration to the changes below (see PyCodeanalyzer._run_analyzer).
Describe the solution you'd like
Not stated in the original issue.
Describe alternatives you've considered
Not stated in the original issue.
Additional context
[1.5.2] - 2026-09-10
Fixed
-
The Neo4j projection now carries the facts analysis.json carries (#202, #203). The two projections are both first-class, so a fact in one and not the other is a bug in the graph: a consumer that ported a JSON traversal to Cypher lost data with no error. Re-expressing a JSON field as an edge or a flattened property is fine; dropping it is not. What was missing, and is now emitted:
:PyModule.source — the module's whole-file text, the anchor every span slices into. Without it a graph consumer could read a node's offsets and had nothing to apply them to.- Full spans everywhere a span is declared. Every label and relationship that carries a span now carries all six properties —
start_line, start_column, end_line, end_column, start_byte, end_byte — matching Span in the payload (start/end as [line, col] plus bytes:[lo, hi]). Previously only the two line numbers reached the graph, so module.source[start_byte:end_byte] had no graph equivalent. The property names are the ones codeanalyzer-java#255 coined, adopted verbatim.
:PyVariable.value_json — the variable's initializer value, JSON-encoded because Neo4j properties are scalars or arrays of scalars and the payload's value is arbitrary (8080, {"a": [1, 2]}). Same treatment arguments_json already had.:PyBodyNode.callee_signature — the call site's resolved callee signature, joined to the body node by position.PY_IMPORTS.positions_json — one position per imported spelling. The edge pre-aggregates per (module, target) and its spellings list is sorted, so the positions are keyed by spelling rather than aligned by index.- A span on
PY_DECORATED_BY — where the decorator is applied. The :PyDecorator node is merged on qualified_name and shared across every use, so the applied position belongs on the relationship, not the node.
One documented exception: :PyAttribute declares a line-only span, because the payload's PyClassAttribute has no column fields to project.
Additive — schema_version and the graph SCHEMA_VERSION both stay 2.0.0 under the hold on the 2.0.0 line; gate on the analyzer version instead. Consumers reading graphs from 1.5.1 or earlier still have to tolerate these properties' absence. Comment nodes remain out of the graph by design and are not part of this fix.
schema.neo4j.json and its docs/handoff/ copy had drifted apart; they are regenerated, identical, and a test now pins them together.
Is your feature request related to a problem? Please describe.
A new
codeanalyzer-pythonrelease (v1.5.2) is published. Update thecodeanalyzer-pythonpin in this repo and adapt the integration to the changes below (seePyCodeanalyzer._run_analyzer).Describe the solution you'd like
Not stated in the original issue.
Describe alternatives you've considered
Not stated in the original issue.
Additional context
[1.5.2] - 2026-09-10
Fixed
The Neo4j projection now carries the facts
analysis.jsoncarries (#202, #203). The two projections are both first-class, so a fact in one and not the other is a bug in the graph: a consumer that ported a JSON traversal to Cypher lost data with no error. Re-expressing a JSON field as an edge or a flattened property is fine; dropping it is not. What was missing, and is now emitted::PyModule.source— the module's whole-file text, the anchor everyspanslices into. Without it a graph consumer could read a node's offsets and had nothing to apply them to.start_line,start_column,end_line,end_column,start_byte,end_byte— matchingSpanin the payload (start/endas[line, col]plusbytes:[lo, hi]). Previously only the two line numbers reached the graph, somodule.source[start_byte:end_byte]had no graph equivalent. The property names are the ones codeanalyzer-java#255 coined, adopted verbatim.:PyVariable.value_json— the variable's initializer value, JSON-encoded because Neo4j properties are scalars or arrays of scalars and the payload'svalueis arbitrary (8080,{"a": [1, 2]}). Same treatmentarguments_jsonalready had.:PyBodyNode.callee_signature— the call site's resolved callee signature, joined to the body node by position.PY_IMPORTS.positions_json— one position per imported spelling. The edge pre-aggregates per(module, target)and itsspellingslist is sorted, so the positions are keyed by spelling rather than aligned by index.PY_DECORATED_BY— where the decorator is applied. The:PyDecoratornode is merged onqualified_nameand shared across every use, so the applied position belongs on the relationship, not the node.One documented exception:
:PyAttributedeclares a line-only span, because the payload'sPyClassAttributehas no column fields to project.Additive —
schema_versionand the graphSCHEMA_VERSIONboth stay2.0.0under the hold on the 2.0.0 line; gate on the analyzer version instead. Consumers reading graphs from 1.5.1 or earlier still have to tolerate these properties' absence. Comment nodes remain out of the graph by design and are not part of this fix.schema.neo4j.jsonand itsdocs/handoff/copy had drifted apart; they are regenerated, identical, and a test now pins them together.