Skip to content

Codebase intelligence for AI and humans: code health scores, auto-generated docs, git analytics, dead code detection, and architectural decisions via MCP.

feat(workspace): bind cross-repo contracts to real symbol ids (#1830)

Raghav Chamadiya82153afb6bb7
diff --git a/packages/core/src/repowise/core/workspace/breaking_change.py b/packages/core/src/repowise/core/workspace/breaking_change.py
index e75c548c8..224b25b8c 100644
--- a/packages/core/src/repowise/core/workspace/breaking_change.py
+++ b/packages/core/src/repowise/core/workspace/breaking_change.py
@@ -87,9 +87,13 @@ class ImpactedConsumer:
symbol: str
match_type: str
confidence: float
+ #: The ingestion symbol id behind ``symbol``, when the contract bound to
+ #: one. ``symbol`` is a display label ("axios:GET /v1/orders"); this is what
+ #: an agent can pass to ``get_symbol`` / ``get_context``.
+ symbol_id: str | None = None
def to_dict(self) -> dict[str, Any]:
- return {
+ d = {
"repo": self.repo,
"service": self.service,
"node_id": self.node_id,
@@ -98,6 +102,9 @@ class ImpactedConsumer:
"match_type": self.match_type,
"confidence": self.confidence,
}
+ if self.symbol_id is not None:
+ d["symbol_id"] = self.symbol_id
+ return d
@classmethod
def from_dict(cls, data: dict[str, Any]) -> ImpactedConsumer:
@@ -109,6 +116,7 @@ class ImpactedConsumer:
symbol=data.get("symbol", ""),
match_type=data.get("match_type", "exact"),
confidence=data.get("confidence", 0.0),
+ symbol_id=data.get("symbol_id"),
)
@@ -125,6 +133,8 @@ class BreakingChange:
provider_symbol: str
provider_service: str | None
detail: str # human-readable one-liner
+ #: The provider handler's symbol id, when its contract bound to one.
+ provider_symbol_id: str | None = None
field_name: str | None = None
old_value: str | None = None
new_value: str | None = None
@@ -148,6 +158,8 @@ class BreakingChange:
"detail": self.detail,
"impacted_consumers": [c.to_dict() for c in self.impacted_consumers],
}
+ if self.provider_symbol_id is not None:
+ d["provider_symbol_id"] = self.provider_symbol_id
if self.field_name is not None:
d["field_name"] = self.field_name
if self.old_value is not None:
@@ -166,6 +178,7 @@ class BreakingChange:
provider_repo=data.get("provider_repo", ""),
provider_file=data.get("provider_file", ""),
provider_symbol=data.get("provider_symbol", ""),
+ provider_symbol_id=data.get("provider_symbol_id"),
provider_service=data.get("provider_service"),
detail=data.get("detail", ""),
field_name=data.get("field_name"),
@@ -486,6 +499,7 @@ def _resolve_impacted(
node_id=node_id,
file=lk.consumer_file,
symbol=lk.consumer_symbol,
+ symbol_id=lk.consumer_symbol_id,
match_type=lk.match_type,
confidence=lk.confidence,
)
@@ -550,6 +564,7 @@ def detect_breaking_changes(
provider_repo=rep.repo,
provider_file=rep.file_path,
provider_symbol=rep.symbol_name,
+ provider_symbol_id=rep.symbol_id,
provider_service=rep.service,
detail=raw.detail,
field_name=raw.field_name,
diff --git a/packages/core/src/repowise/core/workspace/contracts.py b/packages/core/src/repowise/core/workspace/contracts.py
index aa3fb843a..289209296 100644
--- a/packages/core/src/repowise/core/workspace/contracts.py
+++ b/packages/core/src/repowise/core/workspace/contracts.py
@@ -30,7 +30,7 @@ from repowise.core.workspace.contract_schema import ContractSchema
if TYPE_CHECKING:
from repowise.core.workspace.extractors.service_boundary import ServiceBoundary
- from repowise.core.workspace.repo_index import WorkspaceIndex
+ from repowise.core.workspace.repo_index import RepoIndex, WorkspaceIndex
_log = logging.getLogger("repowise.workspace.contracts")
@@ -40,6 +40,11 @@ _log = logging.getLogger("repowise.workspace.contracts")
CONTRACTS_FILENAME = "contracts.json"
+#: Artifact schema version. Bumped to 2 when contracts gained ``symbol_id``.
+#: A store written under an older version is readable but not reusable: its
+#: rows carry no identity, and nothing short of re-extraction can give them one.
+CONTRACTS_VERSION = 2
+
# ---------------------------------------------------------------------------
# Data models
@@ -55,9 +60,18 @@ class Contract:
contract_type: str # "http" | "grpc" | "socket" | "topic" | "data"
role: str # "provider" | "consumer"
file_path: str # relative to repo root
- symbol_name: str # handler name, service.method, etc.
+ symbol_name: str # handler name, service.method, etc. — display only
confidence: float # 0.7–0.9 based on extraction strategy
service: str | None = None # service boundary path (monorepo)
+ #: 1-indexed line of the declaration or call this contract was read from.
+ #: The key :func:`bind_symbol_ids` binds against, and what says *where* a
+ #: contract that failed to bind actually is.
+ line: int | None = None
+ #: The ingestion symbol id (``"<rel_path>::<name>"``) this contract belongs
+ #: to. None when the repo has no index, the file has no parsed symbols, or
+ #: nothing is declared at ``line`` — such a contract still matches, it just
+ #: cannot be traversed into the call graph.
+ symbol_id: str | None = None
meta: dict = field(default_factory=dict)
# Optional request/response shape — populated by dialects that can recover
# it (proto message fields today). Drives schema-level breaking-change diffs.
@@ -67,6 +81,9 @@ class Contract:
d = asdict(self)
if d["service"] is None:
del d["service"]
+ for optional in ("line", "symbol_id"):
+ if d[optional] is None:
+ del d[optional]
if not d["meta"]:
del d["meta"]
if self.schema is None or self.schema.is_empty:
@@ -87,6 +104,9 @@ class Contract:
symbol_name=data["symbol_name"],
confidence=data["confidence"],
service=data.get("service"),
+ # Absent on a v1 artifact: the row predates contract identity.
+ line=data.get("line"),
+ symbol_id=data.get("symbol_id"),
meta=data.get("meta", {}),
schema=ContractSchema.from_dict(raw_schema) if raw_schema else None,
)
@@ -108,6 +128,11 @@ class ContractLink:
consumer_file: str
consumer_symbol: str
consumer_service: str | None
+ #: The linked contracts' symbol ids, carried through so a consumer of this
+ #: link (``ImpactedConsumer``, ``get_risk``) can name the code rather than
+ #: a display label. None when the contract never bound to one.
+ provider_symbol_id: str | None = None
+ consumer_symbol_id: str | None = None
def to_dict(self) -> dict[str, Any]:
d = asdict(self)
@@ -115,6 +140,9 @@ class ContractLink:
del d["provider_service"]
if d["consumer_service"] is None:
del d["consumer_service"]
+ for optional in ("provider_symbol_id", "consumer_symbol_id"):
+ if d[optional] is None:
+ del d[optional]
return d
@classmethod
@@ -132,6 +160,8 @@ class ContractLink:
consumer_file=data["consumer_file"],
consumer_symbol=data.get("consumer_symbol", ""),
consumer_service=data.get("consumer_service"),
+ provider_symbol_id=data.get("provider_symbol_id"),
+ consumer_symbol_id=data.get("consumer_symbol_id"),
)
@@ -139,7 +169,7 @@ class ContractLink:
class ContractStore:
"""Top-level container for contract data, serialized to JSON."""
- version: int = 1
+ version: int = CONTRACTS_VERSION
generated_at: str = ""
contracts: list[Contract] = field(default_factory=list)
contract_links: list[ContractLink] = field(default_factory=list)
@@ -189,6 +219,51 @@ class ContractStore:
return [c for c in self.contracts if c.repo == alias]
+# ---------------------------------------------------------------------------
+# Symbol identity
+# ---------------------------------------------------------------------------
+
+
+# Contract types whose *provider* declaration sits above the symbol it names: a
+# route decorator, an annotation. A data provider is the other shape — a table
+# name is a member inside its owning class — and every consumer is a call inside
+# a body, so for those the symbol containing the line is already the answer and
+# looking below it would take the first method instead.
+_DECLARED_ABOVE = frozenset({"http", "grpc", "socket", "topic"})
+
+
+def bind_symbol_ids(contracts: list[Contract], index: RepoIndex | None) -> dict[str, int]:
+ """Give each contract in *contracts* the id of the symbol it belongs to.
+
+ One pass over every contract type, so identity does not depend on how the
+ dialect found the route — only on the line it reported. Dialects that
+ already know their symbol (the index-backed ones) bind at extraction and
+ are left alone here. Which of the two lookups applies is the one thing a
+ contract's own shape decides: see :data:`_DECLARED_ABOVE`.
+
+ Mutates *contracts* in place and returns the one counter the artifact
+ cannot recover on its own: ``identity_unindexed_<role>``, contracts whose
+ file has no parsed symbols at all — a ``.sql`` file, or a repo with no
+ index. That is the part of the denominator no binding rule can reach, and
+ reporting a ratio without it would blame the rule for it. How many *did*
+ bind is countable from the contracts themselves.
+ """
+ counts: dict[str, int] = {}
+ for contract in contracts:
+ if contract.symbol_id is None and index is not None and contract.line is not None:
+ declares = contract.role == "provider" and contract.contract_type in _DECLARED_ABOVE
+ lookup = index.declared_symbol_at if declares else index.symbol_at
+ symbol = lookup(contract.file_path, contract.line)
+ if symbol is not None:
+ contract.symbol_id = symbol.symbol_id
+ if contract.symbol_id is None and (
+ index is None or not index.symbols_for_file(contract.file_path)
+ ):
+ key = f"identity_unindexed_{contract.role}"
+ counts[key] = counts.get(key, 0) + 1
+ return counts
+
+
# ---------------------------------------------------------------------------
# Contract ID normalization
# ---------------------------------------------------------------------------
@@ -399,6 +474,8 @@ def _make_link(
consumer_repo=consumer.repo,
consumer_file=consumer.file_path,
consumer_symbol=consumer.symbol_name,
+ provider_symbol_id=provider.symbol_id,
+ consumer_symbol_id=consumer.symbol_id,
consumer_service=consumer.service,
)
@@ -830,7 +907,10 @@ async def run_contract_extraction(
json.dumps(contract_config.to_dict(), sort_keys=True).encode()
).hexdigest()[:16]
- if previous_store is not None:
+ # A store written before contracts carried identity holds rows that cannot
+ # be given one without re-reading the source, so its whole reuse question
+ # is moot: extract every repo once, and the next run reuses normally.
+ if previous_store is not None and prior.version >= CONTRACTS_VERSION:
from ..ingestion.change_detector import has_working_tree_changes
aliases = list(repo_paths)
@@ -908,7 +988,8 @@ async def run_contract_extraction(
# The symbols ingestion persisted for this repo. None means the regex
# dialects are the only path, which is also the answer for a repo that
- # has never been indexed. Only the HTTP extractor consults it.
+ # has never been indexed. The HTTP extractor reads it during
+ # extraction; every contract type reads it again in bind_symbol_ids.
repo_index = workspace_index.get(alias) if workspace_index is not None else None
# Counters the HTTP extractor fills in as it goes. The unresolved-path
@@ -937,6 +1018,8 @@ async def run_contract_extraction(
c.meta.setdefault(EXTRACTION_LAYER_KEY, LAYER_REGEX)
contracts.extend(found)
+ stats.update(bind_symbol_ids(contracts, repo_index))
+
unresolved = stats.get("http_consumer_unresolved", 0)
if unresolved:
_log.info(
@@ -990,7 +1073,7 @@ async def run_contract_extraction(
links.extend(_build_manual_links(contract_config.manual_links))
store = ContractStore(
- version=1,
+ version=CONTRACTS_VERSION,
generated_at=now_iso,
contracts=all_contracts,
contract_links=links,
diff --git a/packages/core/src/repowise/core/workspace/diagnostics.py b/packages/core/src/repowise/core/workspace/diagnostics.py
index 25c410d91..566c94a94 100644
--- a/packages/core/src/repowise/core/workspace/diagnostics.py
+++ b/packages/core/src/repowise/core/workspace/diagnostics.py
@@ -109,6 +109,34 @@ class OrphanProvider:
return asdict(self)
+@dataclass
+class SymbolIdentity:
+ """How many of one role's contracts bound to a symbol id, and out of what.
+
+ Two denominators because they answer different questions. *total* is the
+ honest workspace-wide share. *bound_ratio_indexed* excludes the files with
+ no parsed symbols at all — ``.sql``, ``.proto``, anything in a repo without
+ an index — which is what says whether the binding rule itself is working:
+ no rule can reach a file the parser never saw.
+ """
+
+ total: int = 0
+ bound: int = 0
+ unindexed_file: int = 0
+
+ def _ratio(self, denominator: int) -> float | None:
+ return self.bound / denominator if denominator > 0 else None
+
+ def to_dict(self) -> dict[str, Any]:
+ return {
+ "total": self.total,
+ "bound": self.bound,
+ "unindexed_file": self.unindexed_file,
+ "bound_ratio": self._ratio(self.total),
+ "bound_ratio_indexed": self._ratio(self.total - self.unindexed_file),
+ }
+
+
@dataclass
class ExtractionDiagnostics:
"""Aggregate explanation of contract extraction + matching coverage."""
@@ -127,6 +155,9 @@ class ExtractionDiagnostics:
#: HTTP client calls located but not resolvable to an endpoint, workspace
#: wide. This is the only miss count extraction can actually observe.
http_consumers_unresolved: int = 0
+ #: Symbol-id binding coverage, keyed by role. Reported, never asserted: a
+ #: contract with no symbol id still matches, it just cannot be traversed.
+ symbol_identity: dict[str, SymbolIdentity] = field(default_factory=dict)
@property
def http_consumer_coverage(self) -> float | None:
@@ -158,6 +189,7 @@ class ExtractionDiagnostics:
"consumers_by_layer": self.consumers_by_layer,
"http_consumers_unresolved": self.http_consumers_unresolved,
"http_consumer_coverage": self.http_consumer_coverage,
+ "symbol_identity": {r: v.to_dict() for r, v in sorted(self.symbol_identity.items())},
}
@classmethod
@@ -203,6 +235,14 @@ class ExtractionDiagnostics:
providers_by_layer=data.get("providers_by_layer", {}),
consumers_by_layer=data.get("consumers_by_layer", {}),
http_consumers_unresolved=data.get("http_consumers_unresolved", 0),
+ symbol_identity={
+ role: SymbolIdentity(
+ total=v.get("total", 0),
+ bound=v.get("bound", 0),
+ unindexed_file=v.get("unindexed_file", 0),
+ )
+ for role, v in (data.get("symbol_identity") or {}).items()
+ },
)
@@ -338,6 +378,17 @@ def build_diagnostics(
)
)
+ identity = {
+ role: SymbolIdentity(
+ total=len(rows),
+ bound=sum(1 for c in rows if c.symbol_id is not None),
+ unindexed_file=sum(
+ s.get(f"identity_unindexed_{role}", 0) for s in stats_by_repo.values()
+ ),
+ )
+ for role, rows in (("provider", providers), ("consumer", consumers))
+ }
+
return ExtractionDiagnostics(
total_providers=len(providers),
total_consumers=len(consumers),
@@ -355,4 +406,5 @@ def build_diagnostics(
http_consumers_unresolved=sum(
s.get("http_consumer_unresolved", 0) for s in stats_by_repo.values()
),
+ symbol_identity=identity,
)
diff --git a/packages/core/src/repowise/core/workspace/extractors/base.py b/packages/core/src/repowise/core/workspace/extractors/base.py
index 2dd675417..da1e64396 100644
--- a/packages/core/src/repowise/core/workspace/extractors/base.py
+++ b/packages/core/src/repowise/core/workspace/extractors/base.py
@@ -52,6 +52,15 @@ _TEST_FILE_PATTERNS = (
_SELF_EXCLUDE_GLOBS = ("*/repowise/core/workspace/extractors/*",)
+def line_at(content: str, offset: int) -> int:
+ """The 1-indexed line of *content* holding character *offset*.
+
+ Every regex dialect reports where it matched this way, which is what binds
+ a contract to a symbol id (see :func:`..contracts.bind_symbol_ids`).
+ """
+ return content.count("\n", 0, offset) + 1
+
+
def is_test_path(rel_path: str) -> bool:
"""True when *rel_path* (POSIX) lives in a test tree or is a test file."""
parts = rel_path.split("/")
diff --git a/packages/core/src/repowise/core/workspace/extractors/data/dialect.py b/packages/core/src/repowise/core/workspace/extractors/data/dialect.py
index 7e6fd7175..7c5f4b3a8 100644
--- a/packages/core/src/repowise/core/workspace/extractors/data/dialect.py
+++ b/packages/core/src/repowise/core/workspace/extractors/data/dialect.py
@@ -35,6 +35,7 @@ def build_table_provider(
*,
table_raw: str,
framework: str,
+ line: int | None = None,
confidence: float = 0.85,
) -> Contract | None:
"""Build a provider contract for a declared table, or ``None``.
@@ -57,6 +58,7 @@ def build_table_provider(
symbol_name=f"{framework}:{table_raw}",
confidence=confidence,
service=None,
+ line=line,
meta={"table": table, "framework": framework},
)
@@ -67,6 +69,7 @@ def build_table_consumer(
table_raw: str,
verb: str,
client: str,
+ line: int | None = None,
confidence: float = 0.7,
) -> Contract | None:
"""Build a consumer contract for a table referenced from app code.
@@ -89,5 +92,6 @@ def build_table_consumer(
symbol_name=f"{client}:{verb} {table}",
confidence=confidence,
service=None,
+ line=line,
meta={"table": table, "verb": verb, "client": client},
)
diff --git a/packages/core/src/repowise/core/workspace/extractors/data/orm_models.py b/packages/core/src/repowise/core/workspace/extractors/data/orm_models.py
index 2ade4aaba..a4049fee0 100644
--- a/packages/core/src/repowise/core/workspace/extractors/data/orm_models.py
+++ b/packages/core/src/repowise/core/workspace/extractors/data/orm_models.py
@@ -15,9 +15,10 @@ property) are emitted at lower confidence: a wrong guess costs a missed link
from __future__ import annotations
import re
+from collections.abc import Callable
from typing import TYPE_CHECKING
-from ..base import ScanContext
+from ..base import ScanContext, line_at
from ..langs import CSHARP, JAVA, KOTLIN, PHP, PYTHON, RUBY
from .dialect import build_table_provider
@@ -46,19 +47,42 @@ def _pluralize(name: str) -> str:
return name + "s"
+#: One table declaration: ``(raw_name, confidence, line)``.
+_Found = tuple[str, float, int]
+
+
+def _found(
+ pattern: re.Pattern[str],
+ content: str,
+ confidence: float,
+ transform: Callable[[str], str] = str,
+) -> list[_Found]:
+ r"""Every match of *pattern*'s first group, with the line it sits on.
+
+ The line comes from the group, not the match: three of these patterns open
+ with ``^\s*`` under ``MULTILINE``, where ``^`` matches at a preceding blank
+ line and ``\s*`` eats the newlines, putting ``m.start()`` above the
+ declaration.
+ """
+ return [
+ (transform(m.group(1)), confidence, line_at(content, m.start(1)))
+ for m in pattern.finditer(content)
+ ]
+
+
def _dedup_emit(
ctx: ScanContext,
framework: str,
- found: list[tuple[str, float]],
+ found: list[_Found],
) -> list[Contract]:
out: list[Contract] = []
seen: set[str] = set()
- for raw, confidence in found:
+ for raw, confidence, line in found:
if raw in seen:
continue
seen.add(raw)
contract = build_table_provider(
- ctx, table_raw=raw, framework=framework, confidence=confidence
+ ctx, table_raw=raw, framework=framework, line=line, confidence=confidence
)
if contract is not None:
out.append(contract)
@@ -79,12 +103,10 @@ class SqlAlchemyDjangoDialect:
_SQLMODEL_RE = re.compile(r"\bclass\s+(\w+)\([^)]*\btable\s*=\s*True[^)]*\)")
def extract(self, ctx: ScanContext) -> list[Contract]:
- found = [(m, _EXPLICIT_CONFIDENCE) for m in self._TABLENAME_RE.findall(ctx.content)]
- found += [(m, _EXPLICIT_CONFIDENCE) for m in self._DB_TABLE_RE.findall(ctx.content)]
- found += [(m, _EXPLICIT_CONFIDENCE) for m in self._CREATE_TABLE_RE.findall(ctx.content)]
- found += [
- (m.lower(), _CONVENTION_CONFIDENCE) for m in self._SQLMODEL_RE.findall(ctx.content)
- ]
+ found = _found(self._TABLENAME_RE, ctx.content, _EXPLICIT_CONFIDENCE)
+ found += _found(self._DB_TABLE_RE, ctx.content, _EXPLICIT_CONFIDENCE)
+ found += _found(self._CREATE_TABLE_RE, ctx.content, _EXPLICIT_CONFIDENCE)
+ found += _found(self._SQLMODEL_RE, ctx.content, _CONVENTION_CONFIDENCE, str.lower)
return _dedup_emit(ctx, self.name, found)
@@ -101,14 +123,13 @@ class JpaDialect:
)
def extract(self, ctx: ScanContext) -> list[Contract]:
- found = [(m, _EXPLICIT_CONFIDENCE) for m in self._TABLE_RE.findall(ctx.content)]
+ found = _found(self._TABLE_RE, ctx.content, _EXPLICIT_CONFIDENCE)
if not found:
# Only fall back to the class-name convention when no explicit
# @Table names the table (one entity per file is the JPA norm).
- found = [
- (_snake_case(m), _CONVENTION_CONFIDENCE)
- for m in self._ENTITY_CLASS_RE.findall(ctx.content)
- ]
+ found = _found(
+ self._ENTITY_CLASS_RE, ctx.content, _CONVENTION_CONFIDENCE, _snake_case
+ )
return _dedup_emit(ctx, self.name, found)
@@ -122,8 +143,8 @@ class EfCoreDialect:
_DBSET_RE = re.compile(r"\bDbSet<\s*\w+\s*>\s+(\w+)")
def extract(self, ctx: ScanContext) -> list[Contract]:
- found = [(m, _EXPLICIT_CONFIDENCE) for m in self._TABLE_ATTR_RE.findall(ctx.content)]
- found += [(m, _CONVENTION_CONFIDENCE) for m in self._DBSET_RE.findall(ctx.content)]
+ found = _found(self._TABLE_ATTR_RE, ctx.content, _EXPLICIT_CONFIDENCE)
+ found += _found(self._DBSET_RE, ctx.content, _CONVENTION_CONFIDENCE)
return _dedup_emit(ctx, self.name, found)
@@ -139,12 +160,14 @@ class ActiveRecordDialect:
)
def extract(self, ctx: ScanContext) -> list[Contract]:
- found = [(m, _EXPLICIT_CONFIDENCE) for m in self._TABLE_NAME_RE.findall(ctx.content)]
+ found = _found(self._TABLE_NAME_RE, ctx.content, _EXPLICIT_CONFIDENCE)
if not found:
- found = [
- (_pluralize(_snake_case(m)), _CONVENTION_CONFIDENCE)
- for m in self._MODEL_CLASS_RE.findall(ctx.content)
- ]
+ found = _found(
+ self._MODEL_CLASS_RE,
+ ctx.content,
+ _CONVENTION_CONFIDENCE,
+ lambda m: _pluralize(_snake_case(m)),
+ )
return _dedup_emit(ctx, self.name, found)
@@ -157,5 +180,5 @@ class EloquentDialect:
_TABLE_RE = re.compile(r"protected\s+\$table\s*=\s*['\"](\w+)['\"]")
def extract(self, ctx: ScanContext) -> list[Contract]:
- found = [(m, _EXPLICIT_CONFIDENCE) for m in self._TABLE_RE.findall(ctx.content)]
+ found = _found(self._TABLE_RE, ctx.content, _EXPLICIT_CONFIDENCE)
return _dedup_emit(ctx, self.name, found)
diff --git a/packages/core/src/repowise/core/workspace/extractors/data/sql_strings.py b/packages/core/src/repowise/core/workspace/extractors/data/sql_strings.py
index 695ace2b1..d9d9ce111 100644
--- a/packages/core/src/repowise/core/workspace/extractors/data/sql_strings.py
+++ b/packages/core/src/repowise/core/workspace/extractors/data/sql_strings.py
@@ -19,7 +19,7 @@ import logging
import re
from typing import TYPE_CHECKING
-from ..base import ScanContext
+from ..base import ScanContext, line_at
from ..langs import CSHARP, GO, JAVA, JS_TS, KOTLIN, PHP, PYTHON, RUBY
from .dialect import build_table_consumer
@@ -126,7 +126,7 @@ class SqlStringsDialect:
def extract(self, ctx: ScanContext) -> list[Contract]:
out: list[Contract] = []
seen: set[str] = set()
- for literal in self._sql_literals(ctx.content):
+ for literal, line in self._sql_literals(ctx.content):
for table_raw, verb in self._tables_in(literal):
if table_raw.lower().strip('"`[]') in _SQL_KEYWORDS:
continue
@@ -135,15 +135,16 @@ class SqlStringsDialect:
continue
seen.add(key)
contract = build_table_consumer(
- ctx, table_raw=table_raw, verb=verb, client=self.name
+ ctx, table_raw=table_raw, verb=verb, client=self.name, line=line
)
if contract is not None:
out.append(contract)
return out
@staticmethod
- def _sql_literals(content: str) -> list[str]:
- literals: list[str] = []
+ def _sql_literals(content: str) -> list[tuple[str, int]]:
+ """``(literal, line)`` for each SQL-looking string literal in *content*."""
+ literals: list[tuple[str, int]] = []
for m in _STRING_RE.finditer(content):
s = next(g for g in m.groups() if g is not None)
if (
@@ -152,7 +153,7 @@ class SqlStringsDialect:
and _SQL_VERB_RE.search(s)
and (_UPPER_VERB_RE.search(s) or _SQL_PUNCT_RE.search(s))
):
- literals.append(s)
+ literals.append((s, line_at(content, m.start())))
return literals
def _tables_in(self, literal: str) -> list[tuple[str, str]]:
diff --git a/packages/core/src/repowise/core/workspace/extractors/from_index.py b/packages/core/src/repowise/core/workspace/extractors/from_index.py
index bf947c177..a431b5416 100644
--- a/packages/core/src/repowise/core/workspace/extractors/from_index.py
+++ b/packages/core/src/repowise/core/workspace/extractors/from_index.py
@@ -273,11 +273,15 @@ def extract_http_providers(
seen.add(key)
framework = "flask" if _is_route_head(decorator) else "fastapi"
contract = build_provider_contract(
- ctx, method=method, path_raw=path, framework=framework
+ ctx, method=method, path_raw=path, framework=framework, line=symbol.start_line
)
if contract is not None:
contract.meta[EXTRACTION_LAYER_KEY] = LAYER_INDEX
contract.meta["handler"] = symbol.name
+ # Bound here rather than by line: this path found the route *by*
+ # walking up from the handler's own span, so the symbol is known
+ # exactly and no lookahead rule needs to re-derive it.
+ contract.symbol_id = symbol.symbol_id
out.append(contract)
return out
diff --git a/packages/core/src/repowise/core/workspace/extractors/grpc/csharp.py b/packages/core/src/repowise/core/workspace/extractors/grpc/csharp.py
index 993ff2d57..68d189b1f 100644
--- a/packages/core/src/repowise/core/workspace/extractors/grpc/csharp.py
+++ b/packages/core/src/repowise/core/workspace/extractors/grpc/csharp.py
@@ -13,6 +13,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import CSHARP
from .dialect import make_grpc_contract
@@ -55,6 +56,7 @@ class CSharpGrpcDialect:
symbol_name=f"cs:MapGrpcService<{svc}>",
confidence=0.8,
meta={"service": svc, "source": "csharp_mapgrpc"},
+ line=line_at(ctx.content, m.start()),
)
)
for m in _CSHARP_GRPC_BASE_RE.finditer(ctx.content):
@@ -67,6 +69,7 @@ class CSharpGrpcDialect:
symbol_name=f"cs:extends {svc}.{svc}ServiceBase",
confidence=0.8,
meta={"service": svc, "source": "csharp_base"},
+ line=line_at(ctx.content, m.start()),
)
)
# Consumers are only credible when the file shows real gRPC usage —
@@ -85,6 +88,7 @@ class CSharpGrpcDialect:
symbol_name=f"cs:new {svc}Client",
confidence=0.65,
meta={"service": svc, "source": "csharp_client"},
+ line=line_at(ctx.content, m.start()),
)
)
return out
diff --git a/packages/core/src/repowise/core/workspace/extractors/grpc/dialect.py b/packages/core/src/repowise/core/workspace/extractors/grpc/dialect.py
index 2685892fc..f727c1a10 100644
--- a/packages/core/src/repowise/core/workspace/extractors/grpc/dialect.py
+++ b/packages/core/src/repowise/core/workspace/extractors/grpc/dialect.py
@@ -35,6 +35,7 @@ def make_grpc_contract(
symbol_name: str,
confidence: float,
meta: dict,
+ line: int | None = None,
) -> Contract:
"""Build a gRPC :class:`Contract` with the common fields filled in."""
from repowise.core.workspace.contracts import Contract
@@ -48,5 +49,6 @@ def make_grpc_contract(
symbol_name=symbol_name,
confidence=confidence,
service=None,
+ line=line,
meta=meta,
)
diff --git a/packages/core/src/repowise/core/workspace/extractors/grpc/go.py b/packages/core/src/repowise/core/workspace/extractors/grpc/go.py
index ec89cdd6f..509d5edfe 100644
--- a/packages/core/src/repowise/core/workspace/extractors/grpc/go.py
+++ b/packages/core/src/repowise/core/workspace/extractors/grpc/go.py
@@ -5,6 +5,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import GO
from .dialect import make_grpc_contract
@@ -35,6 +36,7 @@ class GoGrpcDialect:
symbol_name=f"go:Register{svc}Server",
confidence=0.8,
meta={"service": svc, "source": "go_register"},
+ line=line_at(ctx.content, m.start()),
)
)
for m in _GO_CONSUMER_RE.finditer(ctx.content):
@@ -47,6 +49,7 @@ class GoGrpcDialect:
symbol_name=f"go:New{svc}Client",
confidence=0.7,
meta={"service": svc, "source": "go_client"},
+ line=line_at(ctx.content, m.start()),
)
)
return out
diff --git a/packages/core/src/repowise/core/workspace/extractors/grpc/java.py b/packages/core/src/repowise/core/workspace/extractors/grpc/java.py
index 4402b2fda..5dfef7bb6 100644
--- a/packages/core/src/repowise/core/workspace/extractors/grpc/java.py
+++ b/packages/core/src/repowise/core/workspace/extractors/grpc/java.py
@@ -5,6 +5,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import JAVA
from .dialect import make_grpc_contract
@@ -35,6 +36,7 @@ class JavaGrpcDialect:
symbol_name=f"java:extends {svc}Grpc.ImplBase",
confidence=0.8,
meta={"service": svc, "source": "java_extends"},
+ line=line_at(ctx.content, m.start()),
)
)
for m in _JAVA_CONSUMER_RE.finditer(ctx.content):
@@ -47,6 +49,7 @@ class JavaGrpcDialect:
symbol_name=f"java:{svc}Grpc.newStub",
confidence=0.7,
meta={"service": svc, "source": "java_stub"},
+ line=line_at(ctx.content, m.start()),
)
)
return out
diff --git a/packages/core/src/repowise/core/workspace/extractors/grpc/python.py b/packages/core/src/repowise/core/workspace/extractors/grpc/python.py
index cdf6b0586..1a0378571 100644
--- a/packages/core/src/repowise/core/workspace/extractors/grpc/python.py
+++ b/packages/core/src/repowise/core/workspace/extractors/grpc/python.py
@@ -5,6 +5,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import PYTHON
from .dialect import make_grpc_contract
@@ -38,6 +39,7 @@ class PythonGrpcDialect:
symbol_name=f"py:add_{svc}Servicer_to_server",
confidence=0.8,
meta={"service": svc, "source": "py_servicer"},
+ line=line_at(ctx.content, m.start()),
)
)
for m in _PY_CONSUMER_RE.finditer(ctx.content):
@@ -52,6 +54,7 @@ class PythonGrpcDialect:
symbol_name=f"py:{svc}Stub",
confidence=0.7,
meta={"service": svc, "source": "py_stub"},
+ line=line_at(ctx.content, m.start()),
)
)
return out
diff --git a/packages/core/src/repowise/core/workspace/extractors/grpc/typescript.py b/packages/core/src/repowise/core/workspace/extractors/grpc/typescript.py
index b617fd4f6..2b1eefe43 100644
--- a/packages/core/src/repowise/core/workspace/extractors/grpc/typescript.py
+++ b/packages/core/src/repowise/core/workspace/extractors/grpc/typescript.py
@@ -5,6 +5,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import JS_TS
from .dialect import make_grpc_contract
@@ -34,6 +35,7 @@ class TypeScriptGrpcDialect:
symbol_name=f"ts:@GrpcMethod('{svc}', '{method}')",
confidence=0.8,
meta={"service": svc, "method": method, "source": "ts_decorator"},
+ line=line_at(ctx.content, m.start()),
)
)
return out
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/aspnet.py b/packages/core/src/repowise/core/workspace/extractors/http/aspnet.py
index 5bf38696a..48f433444 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/aspnet.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/aspnet.py
@@ -14,6 +14,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import CSHARP
from .dialect import METHODS, build_provider_contract, nearest_prefix
@@ -67,14 +68,24 @@ class AspNetDialect:
prefix = nearest_prefix(class_mappings, m.start())
if prefix:
path_raw = prefix + "/" + path_raw.lstrip("/")
- c = build_provider_contract(ctx, method=method, path_raw=path_raw, framework="aspnet")
+ c = build_provider_contract(
+ ctx,
+ method=method,
+ path_raw=path_raw,
+ framework="aspnet",
+ line=line_at(content, m.start()),
+ )
if c is not None:
out.append(c)
# Minimal API — not inside a controller, never prefix-stitched.
for m in _ASPNET_MINIMAL_RE.finditer(content):
c = build_provider_contract(
- ctx, method=m.group(1).upper(), path_raw=m.group(2), framework="aspnet-minimal"
+ ctx,
+ method=m.group(1).upper(),
+ path_raw=m.group(2),
+ framework="aspnet-minimal",
+ line=line_at(content, m.start()),
)
if c is not None:
out.append(c)
@@ -87,7 +98,13 @@ class AspNetDialect:
prefix = nearest_prefix(class_mappings, m.start())
if not prefix:
continue
- c = build_provider_contract(ctx, method=method, path_raw=prefix, framework="aspnet")
+ c = build_provider_contract(
+ ctx,
+ method=method,
+ path_raw=prefix,
+ framework="aspnet",
+ line=line_at(content, m.start()),
+ )
if c is not None:
out.append(c)
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/csharp_http.py b/packages/core/src/repowise/core/workspace/extractors/http/csharp_http.py
index 690e3e8d9..0dc1e4a8c 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/csharp_http.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/csharp_http.py
@@ -18,6 +18,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import CSHARP
from .dialect import build_consumer_contract
@@ -66,22 +67,22 @@ class CSharpHttpDialect:
content = ctx.content
out: list[Contract] = []
- def emit(method: str, prefix: str, text: str, client: str) -> None:
+ def emit(method: str, prefix: str, text: str, client: str, line: int) -> None:
url = _to_template(prefix, text)
if "/" not in url:
return # Not a URL path — skip non-route strings.
c = build_consumer_contract(
- ctx, method=method.upper(), url=url, client=client, confidence=0.70
+ ctx, method=method.upper(), url=url, client=client, line=line, confidence=0.70
)
if c is not None:
out.append(c)
for m in _WRAPPER_RE.finditer(content):
- emit(m.group(1), m.group(2), m.group(3), "httpclient")
+ emit(m.group(1), m.group(2), m.group(3), "httpclient", line_at(content, m.start()))
for m in _UNITY_RE.finditer(content):
- emit(m.group(1), m.group(2), m.group(3), "unitywebrequest")
+ emit(m.group(1), m.group(2), m.group(3), "unitywebrequest", line_at(content, m.start()))
for m in _BESTHTTP_RE.finditer(content):
# Group order: prefix, text, method (method trails the URL here).
- emit(m.group(3), m.group(1), m.group(2), "besthttp")
+ emit(m.group(3), m.group(1), m.group(2), "besthttp", line_at(content, m.start()))
return out
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/dialect.py b/packages/core/src/repowise/core/workspace/extractors/http/dialect.py
index 18606dd7f..56aded7a0 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/dialect.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/dialect.py
@@ -64,6 +64,7 @@ def build_provider_contract(
method: str,
path_raw: str,
framework: str,
+ line: int | None = None,
confidence: float = 0.85,
) -> Contract | None:
"""Build a provider contract, or ``None`` if the path is unusable.
@@ -87,6 +88,7 @@ def build_provider_contract(
symbol_name=f"{framework}:{method} {path_raw}",
confidence=confidence,
service=None,
+ line=line,
meta={"method": method, "path": norm_path, "framework": framework},
)
@@ -97,6 +99,7 @@ def build_consumer_contract(
method: str,
url: str,
client: str,
+ line: int | None = None,
confidence: float = 0.75,
) -> Contract | None:
"""Build a consumer contract from a raw client-call URL.
@@ -122,5 +125,6 @@ def build_consumer_contract(
symbol_name=f"{client}:{method} {url}",
confidence=confidence,
service=None,
+ line=line,
meta=consumer_meta(method, norm_path, client, base_token, host),
)
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/express.py b/packages/core/src/repowise/core/workspace/extractors/http/express.py
index 501c2bda8..16c37ac77 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/express.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/express.py
@@ -10,6 +10,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import JS_TS
from .dialect import METHODS, build_provider_contract
from .mounts import compose_prefix
@@ -59,7 +60,11 @@ class ExpressDialect:
continue
path = compose_prefix(ctx.mounts.get(var, ""), path_raw)
c = build_provider_contract(
- ctx, method=method.upper(), path_raw=path, framework="express"
+ ctx,
+ method=method.upper(),
+ path_raw=path,
+ framework="express",
+ line=line_at(ctx.content, m.start()),
)
if c is not None:
out.append(c)
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/fastapi.py b/packages/core/src/repowise/core/workspace/extractors/http/fastapi.py
index 3b37f6ad5..349aa089c 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/fastapi.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/fastapi.py
@@ -10,6 +10,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import PYTHON
from .dialect import METHODS, build_provider_contract
from .mounts import balanced_args, compose_prefix, router_prefixes
@@ -72,7 +73,11 @@ class FastApiDialect:
path = compose_prefix(prefixes.get(var, ""), path_raw)
path = compose_prefix(ctx.mounts.get(var, ""), path)
c = build_provider_contract(
- ctx, method=method.upper(), path_raw=path, framework="fastapi"
+ ctx,
+ method=method.upper(),
+ path_raw=path,
+ framework="fastapi",
+ line=line_at(ctx.content, m.start()),
)
if c is not None:
out.append(c)
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/go.py b/packages/core/src/repowise/core/workspace/extractors/http/go.py
index 304bae987..be2ad7d5b 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/go.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/go.py
@@ -11,6 +11,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import GO
from .dialect import METHODS_UPPER, build_provider_contract
from .mounts import compose_prefix
@@ -68,7 +69,13 @@ class GoDialect:
# Handle/HandleFunc don't carry a method verb.
method = "*" if method_raw in ("Handle", "HandleFunc") else method_raw.upper()
path = compose_prefix(prefixes.get(var, ""), path_raw)
- c = build_provider_contract(ctx, method=method, path_raw=path, framework="go")
+ c = build_provider_contract(
+ ctx,
+ method=method,
+ path_raw=path,
+ framework="go",
+ line=line_at(ctx.content, m.start()),
+ )
if c is not None:
out.append(c)
return out
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/index_clients.py b/packages/core/src/repowise/core/workspace/extractors/http/index_clients.py
index 06ab8d8a3..dfa8bbe81 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/index_clients.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/index_clients.py
@@ -21,6 +21,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from .dialect import build_consumer_contract
from .wrappers import (
DEFAULT_HOP_BUDGET,
@@ -230,7 +231,7 @@ def extract_consumers(
# (concrete literal) and unresolved. The set of wrappers proven to take a
# path is only complete once the whole file has been scanned, so the
# unresolved tally is settled afterwards.
- resolved: list[tuple[str, str, str]] = [] # (callee, url, method)
+ resolved: list[tuple[str, str, str, int]] = [] # (callee, url, method, line)
unresolved_by_callee: dict[str, int] = {}
path_taking: set[str] = set()
# Calls whose argument list would not parse. Kept apart from the
@@ -251,7 +252,7 @@ def extract_consumers(
is_sink_call = m.group("recv") is None and name in sinks
if not is_sink_call and name not in confirmed:
continue
- line = content.count("\n", 0, m.start()) + 1
+ line = line_at(content, m.start())
if (line, name) in declarations:
continue # the wrapper's own signature, not a call to it
open_idx = m.end() - 1
@@ -276,17 +277,17 @@ def extract_consumers(
continue
path_taking.add(name)
opt = _METHOD_OPT_RE.search(rest)
- resolved.append((name, url, opt.group(1).upper() if opt else "GET"))
+ resolved.append((name, url, opt.group(1).upper() if opt else "GET", line))
from ..from_index import EXTRACTION_LAYER_KEY, LAYER_INDEX
out: list[Contract] = []
- for callee, url, method in resolved:
+ for callee, url, method, line in resolved:
# Higher than the regex dialect's 0.65/0.75: the callee is a confirmed
# sink-reaching wrapper and the URL is a whole literal argument, not a
# string that happened to sit next to a parenthesis.
contract = build_consumer_contract(
- ctx, method=method, url=url, client=callee, confidence=0.85
+ ctx, method=method, url=url, client=callee, line=line, confidence=0.85
)
if contract is not None:
contract.meta[EXTRACTION_LAYER_KEY] = LAYER_INDEX
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/js_clients.py b/packages/core/src/repowise/core/workspace/extractors/http/js_clients.py
index 67f84f5a3..e2175d9de 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/js_clients.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/js_clients.py
@@ -12,6 +12,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import JS_TS
from .dialect import METHODS, build_consumer_contract
@@ -59,16 +60,23 @@ class JsClientsDialect:
content = ctx.content
out: list[Contract] = []
- def emit(*, method: str, url: str, client: str, confidence: float = 0.75) -> None:
+ def emit(
+ *, method: str, url: str, client: str, line: int, confidence: float = 0.75
+ ) -> None:
c = build_consumer_contract(
- ctx, method=method, url=url, client=client, confidence=confidence
+ ctx, method=method, url=url, client=client, line=line, confidence=confidence
)
if c is not None:
out.append(c)
# fetch() with an explicit method.
for m in _FETCH_METHOD_RE.finditer(content):
- emit(method=m.group(2).upper(), url=m.group(1), client="fetch")
+ emit(
+ method=m.group(2).upper(),
+ url=m.group(1),
+ client="fetch",
+ line=line_at(content, m.start()),
+ )
# fetch() without a method → GET, skipping URLs already matched above.
method_urls = {m.group(1) for m in _FETCH_METHOD_RE.finditer(content)}
@@ -76,11 +84,16 @@ class JsClientsDialect:
url = m.group(1)
if url in method_urls:
continue
- emit(method="GET", url=url, client="fetch")
+ emit(method="GET", url=url, client="fetch", line=line_at(content, m.start()))
# axios.<method>()
for m in _AXIOS_RE.finditer(content):
- emit(method=m.group(1).upper(), url=m.group(2), client="axios")
+ emit(
+ method=m.group(1).upper(),
+ url=m.group(2),
+ client="axios",
+ line=line_at(content, m.start()),
+ )
# Wrapper calls — fetchJSON(`${BASE}/path`, { method: "POST" }) etc.
for m in _WRAPPER_CALL_RE.finditer(content):
@@ -94,6 +107,12 @@ class JsClientsDialect:
if not (_HTTP_NAME_RE.search(callee) or method_opt):
continue
method = method_opt.group(1).upper() if method_opt else "GET"
- emit(method=method, url=m.group(2), client="wrapper", confidence=0.65)
+ emit(
+ method=method,
+ url=m.group(2),
+ client="wrapper",
+ line=line_at(content, m.start()),
+ confidence=0.65,
+ )
return out
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/laravel.py b/packages/core/src/repowise/core/workspace/extractors/http/laravel.py
index f4a40a2da..117ef6e07 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/laravel.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/laravel.py
@@ -5,6 +5,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import PHP
from .dialect import METHODS, build_provider_contract
@@ -28,7 +29,11 @@ class LaravelDialect:
out: list[Contract] = []
for m in _LARAVEL_RE.finditer(ctx.content):
c = build_provider_contract(
- ctx, method=m.group(1).upper(), path_raw=m.group(2), framework="laravel"
+ ctx,
+ method=m.group(1).upper(),
+ path_raw=m.group(2),
+ framework="laravel",
+ line=line_at(ctx.content, m.start()),
)
if c is not None:
out.append(c)
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/python_clients.py b/packages/core/src/repowise/core/workspace/extractors/http/python_clients.py
index 52302617d..fa4cba52b 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/python_clients.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/python_clients.py
@@ -5,6 +5,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import PYTHON
from .dialect import METHODS, build_consumer_contract
@@ -28,7 +29,11 @@ class PythonClientsDialect:
out: list[Contract] = []
for m in _REQUESTS_RE.finditer(ctx.content):
c = build_consumer_contract(
- ctx, method=m.group(1).upper(), url=m.group(2), client="requests"
+ ctx,
+ method=m.group(1).upper(),
+ url=m.group(2),
+ client="requests",
+ line=line_at(ctx.content, m.start()),
)
if c is not None:
out.append(c)
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/rust_axum.py b/packages/core/src/repowise/core/workspace/extractors/http/rust_axum.py
index d1c9815a6..823b58099 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/rust_axum.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/rust_axum.py
@@ -15,6 +15,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import RUST
from .dialect import build_provider_contract
@@ -63,7 +64,11 @@ class RustAxumDialect:
methods = {mm.group(1).upper() for mm in _AXUM_METHOD_RE.finditer(window)}
for method in sorted(methods):
c = build_provider_contract(
- ctx, method=method, path_raw=path_raw, framework="axum"
+ ctx,
+ method=method,
+ path_raw=path_raw,
+ framework="axum",
+ line=line_at(content, m.start()),
)
if c is not None:
out.append(c)
@@ -71,7 +76,11 @@ class RustAxumDialect:
# Actix-web / Rocket attribute macros.
for m in _RUST_ATTR_ROUTE_RE.finditer(content):
c = build_provider_contract(
- ctx, method=m.group(1).upper(), path_raw=m.group(2), framework="rust-attr"
+ ctx,
+ method=m.group(1).upper(),
+ path_raw=m.group(2),
+ framework="rust-attr",
+ line=line_at(content, m.start()),
)
if c is not None:
out.append(c)
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/rust_clients.py b/packages/core/src/repowise/core/workspace/extractors/http/rust_clients.py
index 87db3e180..57508bf3b 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/rust_clients.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/rust_clients.py
@@ -16,6 +16,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import RUST
from .dialect import build_consumer_contract
@@ -49,22 +50,22 @@ class RustClientsDialect:
content = ctx.content
out: list[Contract] = []
- def emit(method: str, url: str) -> None:
+ def emit(method: str, url: str, line: int) -> None:
if "/" not in url:
return # Not a URL path — skip map/collection .get("key") calls.
c = build_consumer_contract(
- ctx, method=method.upper(), url=url, client="reqwest", confidence=0.65
+ ctx, method=method.upper(), url=url, client="reqwest", line=line, confidence=0.65
)
if c is not None:
out.append(c)
for rx in (_METHOD_LIT_RE, _FREE_LIT_RE):
for m in rx.finditer(content):
- emit(m.group(1), m.group(2))
+ emit(m.group(1), m.group(2), line_at(content, m.start()))
for rx in (_METHOD_FMT_RE, _FREE_FMT_RE):
for m in rx.finditer(content):
url = _FMT_PLACEHOLDER_RE.sub("${x}", m.group(2))
- emit(m.group(1), url)
+ emit(m.group(1), url, line_at(content, m.start()))
return out
diff --git a/packages/core/src/repowise/core/workspace/extractors/http/spring.py b/packages/core/src/repowise/core/workspace/extractors/http/spring.py
index 71547b8f3..87e16c11c 100644
--- a/packages/core/src/repowise/core/workspace/extractors/http/spring.py
+++ b/packages/core/src/repowise/core/workspace/extractors/http/spring.py
@@ -9,6 +9,7 @@ from __future__ import annotations
import re
from typing import TYPE_CHECKING
+from ..base import line_at
from ..langs import JAVA
from .dialect import build_provider_contract, nearest_prefix
@@ -49,7 +50,13 @@ class SpringDialect:
prefix = nearest_prefix(class_mappings, m.start())
if prefix:
path_raw = prefix + "/" + path_raw.lstrip("/")
- c = build_provider_contract(ctx, method=method, path_raw=path_raw, framework="spring")
+ c = build_provider_contract(
+ ctx,
+ method=method,
+ path_raw=path_raw,
+ framework="spring",
+ line=line_at(content, m.start()),
+ )
if c is not None:
out.append(c)
return out
diff --git a/packages/core/src/repowise/core/workspace/extractors/socket_extractor.py b/packages/core/src/repowise/core/workspace/extractors/socket_extractor.py
index 56b28e4c2..8bab4515e 100644
--- a/packages/core/src/repowise/core/workspace/extractors/socket_extractor.py
+++ b/packages/core/src/repowise/core/workspace/extractors/socket_extractor.py
@@ -12,7 +12,7 @@ from dataclasses import dataclass
from pathlib import Path
from typing import TYPE_CHECKING
-from .base import select_files
+from .base import line_at, select_files
from .http.paths import (
extract_path_from_url,
is_unusable_consumer_path,
@@ -189,6 +189,7 @@ class SocketExtractor:
symbol_name=f"{pdef.label}('{identity}')",
confidence=pdef.confidence,
service=None,
+ line=line_at(content, match.start()),
meta={
"path": identity,
"transport": pdef.transport,
diff --git a/packages/core/src/repowise/core/workspace/extractors/topic_extractor.py b/packages/core/src/repowise/core/workspace/extractors/topic_extractor.py
index a5882dd98..6c272112f 100644
--- a/packages/core/src/repowise/core/workspace/extractors/topic_extractor.py
+++ b/packages/core/src/repowise/core/workspace/extractors/topic_extractor.py
@@ -14,7 +14,7 @@ from typing import TYPE_CHECKING
from repowise.core.ingestion.languages.registry import REGISTRY as _LANG_REGISTRY
-from .base import select_files
+from .base import line_at, select_files
if TYPE_CHECKING:
from collections.abc import Callable, Sequence
@@ -271,6 +271,7 @@ class TopicExtractor:
symbol_name=f"{pdef.label}('{topic_name}')",
confidence=pdef.confidence,
service=None,
+ line=line_at(content, match.start()),
meta={
"topic": topic_name,
"broker": pdef.broker,
diff --git a/packages/core/src/repowise/core/workspace/repo_index.py b/packages/core/src/repowise/core/workspace/repo_index.py
index 7afd5e122..cf7faf778 100644
--- a/packages/core/src/repowise/core/workspace/repo_index.py
+++ b/packages/core/src/repowise/core/workspace/repo_index.py
@@ -33,6 +33,14 @@ _log = logging.getLogger("repowise.workspace.repo_index")
# the repo — a third-party package or a sibling workspace repo.
EXTERNAL_PREFIX = "external:"
+# How far below a declaration line its symbol may open. Covers a stack of
+# decorators or annotations and the blank lines between them. The ceiling: a
+# declaration this close above an unrelated *nested* definition binds to it,
+# since the index alone cannot tell a decorator run from ordinary code. Reading
+# the intervening text would settle it, at the cost of the caller holding the
+# file (see :func:`.extractors.from_index._decorators_above`, which does).
+_DECLARATION_LOOKAHEAD = 8
+
@dataclass(frozen=True)
class IndexedSymbol:
@@ -143,6 +151,36 @@ class RepoIndex:
best = sym
return best
+ def declared_symbol_at(self, rel_path: str, line: int) -> IndexedSymbol | None:
+ """The symbol a declaration on *line* names, or None.
+
+ :meth:`symbol_at` answers "which symbol contains this line", which is
+ the wrong question for a declaration that sits *above* its handler: a
+ route decorator or annotation is outside the handler's span, because
+ the parser takes a symbol's span from the definition node rather than
+ the decorated one. So a symbol opening just below *line* is preferred,
+ but only when it is nested inside whatever contains *line* — that guard
+ is what stops a call on a function's last line from binding to the next
+ function down.
+ """
+ containing = self.symbol_at(rel_path, line)
+ # A match *on* a symbol's own declaration line names that symbol, not
+ # the first member under it (a gRPC servicer class, say).
+ if containing is not None and containing.start_line == line:
+ return containing
+ following: IndexedSymbol | None = None
+ for sym in self._by_file.get(rel_path, ()):
+ if line < sym.start_line <= line + _DECLARATION_LOOKAHEAD and (
+ following is None or sym.start_line < following.start_line
+ ):
+ following = sym
+ if following is not None and (
+ containing is None
+ or containing.start_line <= following.start_line <= containing.end_line
+ ):
+ return following
+ return containing
+
def external_import_edges(self) -> list[ExternalImport]:
"""Every ``imports`` edge leaving the repo, with the names it consumes."""
return self._externals
diff --git a/packages/server/src/repowise/server/mcp_server/tool_risk/directives.py b/packages/server/src/repowise/server/mcp_server/tool_risk/directives.py
index 11a9a16e8..c08540f72 100644
--- a/packages/server/src/repowise/server/mcp_server/tool_risk/directives.py
+++ b/packages/server/src/repowise/server/mcp_server/tool_risk/directives.py
@@ -86,10 +86,14 @@ def _breaking_change_directive(repo_alias: str) -> list[dict[str, Any]]:
"severity": change.get("severity"),
"detail": change.get("detail"),
"impacted_consumers": [
+ # symbol_id only when the contract bound to one: it is
+ # what the reader can pass to get_symbol, and a null
+ # would just cost budget.
{
"repo": c.get("repo"),
"service": c.get("service"),
"file": c.get("file"),
+ **({"symbol_id": sid} if (sid := c.get("symbol_id")) else {}),
}
for c in cross[:_BC_CONSUMER_LIMIT]
],
diff --git a/packages/server/src/repowise/server/schemas/workspace.py b/packages/server/src/repowise/server/schemas/workspace.py
index 1937ee04d..83848b81f 100644
--- a/packages/server/src/repowise/server/schemas/workspace.py
+++ b/packages/server/src/repowise/server/schemas/workspace.py
@@ -274,7 +274,8 @@ class WorkspaceImpactedConsumer(BaseModel):
service: str | None = None
node_id: str
file: str
- symbol: str
+ symbol: str # display label; symbol_id is the one a tool can look up
+ symbol_id: str | None = None
match_type: str = "exact"
confidence: float = 0.0
diff --git a/tests/unit/workspace/test_contract_identity.py b/tests/unit/workspace/test_contract_identity.py
new file mode 100644
index 000000000..53b45466a
--- /dev/null
+++ b/tests/unit/workspace/test_contract_identity.py
@@ -0,0 +1,406 @@
+"""Contract identity: the symbol id a contract belongs to.
+
+A contract's ``symbol_name`` is a display label — ``"axios:GET /v1/orders"`` —
+so it names nothing an agent can look up. ``symbol_id`` is the ingestion symbol
+id, bound from the line the dialect matched on against the repo's symbol table.
+
+Two claims are under test. The first is that binding is language-agnostic: it
+reads a line and a symbol table, so a dialect that never heard of it gets
+identity by reporting where it matched. The second is that it is additive
+everywhere — a contract that cannot bind still matches, and a v1 artifact still
+loads.
+"""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+import pytest
+
+from repowise.core.ingestion.models import Symbol
+from repowise.core.workspace.config import ContractConfig, RepoEntry, WorkspaceConfig
+from repowise.core.workspace.contracts import (
+ CONTRACTS_VERSION,
+ Contract,
+ ContractStore,
+ bind_symbol_ids,
+ run_contract_extraction,
+)
+from repowise.core.workspace.diagnostics import build_diagnostics
+
+from ._repo_index import make_repo_index
+
+
+def _symbol(name: str, start: int, end: int, *, path="app/api.py", kind="function"):
+ return Symbol(
+ id=f"{path}::{name}",
+ name=name,
+ qualified_name=name,
+ kind=kind,
+ signature=f"def {name}()",
+ start_line=start,
+ end_line=end,
+ docstring=None,
+ visibility="public",
+ )
+
+
+def _contract(line: int | None, *, role="provider", path="app/api.py") -> Contract:
+ return Contract(
+ repo="backend",
+ contract_id="http::GET::/users",
+ contract_type="http",
+ role=role,
+ file_path=path,
+ symbol_name="fastapi:GET /users",
+ confidence=0.85,
+ line=line,
+ )
+
+
+# ---------------------------------------------------------------------------
+# Binding by line
+# ---------------------------------------------------------------------------
+
+
+class TestDeclaredSymbolAt:
+ """The lookup that turns a matched line into a symbol.
+
+ ``symbol_at`` alone is not enough: a route decorator sits *above* its
+ handler, outside the span, because the parser takes a symbol's extent from
+ the ``def`` node rather than the decorated one.
+ """
+
+ async def test_a_decorator_binds_to_the_handler_below_it(self, tmp_path: Path) -> None:
+ index = await make_repo_index(tmp_path, {"app/api.py": [_symbol("users", 11, 14)]})
+ try:
+ assert index.symbol_at("app/api.py", 10) is None # outside the span
+ found = index.declared_symbol_at("app/api.py", 10)
+ assert found is not None and found.name == "users"
+ finally:
+ await index.close()
+
+ async def test_a_decorated_method_wins_over_its_class(self, tmp_path: Path) -> None:
+ index = await make_repo_index(
+ tmp_path,
+ {
+ "app/api.py": [
+ _symbol("Views", 1, 20, kind="class"),
+ _symbol("get", 5, 8, kind="method"),
+ ]
+ },
+ )
+ try:
+ assert index.symbol_at("app/api.py", 4).name == "Views"
+ assert index.declared_symbol_at("app/api.py", 4).name == "get"
+ finally:
+ await index.close()
+
+ async def test_a_call_inside_a_body_binds_to_that_body(self, tmp_path: Path) -> None:
+ index = await make_repo_index(tmp_path, {"app/api.py": [_symbol("fetch_all", 1, 9)]})
+ try:
+ assert index.declared_symbol_at("app/api.py", 5).name == "fetch_all"
+ finally:
+ await index.close()
+
+ async def test_a_call_on_a_last_line_does_not_leak_to_the_next_symbol(
+ self, tmp_path: Path
+ ) -> None:
+ """The nesting guard: the next definition down is not this one's."""
+ index = await make_repo_index(
+ tmp_path,
+ {"app/api.py": [_symbol("first", 1, 5), _symbol("second", 7, 12)]},
+ )
+ try:
+ assert index.declared_symbol_at("app/api.py", 5).name == "first"
+ finally:
+ await index.close()
+
+ async def test_a_match_on_a_declaration_line_names_that_symbol(
+ self, tmp_path: Path
+ ) -> None:
+ """A gRPC servicer class, not the first method inside it."""
+ index = await make_repo_index(
+ tmp_path,
+ {
+ "app/api.py": [
+ _symbol("OrderService", 2, 30, kind="class"),
+ _symbol("get_order", 5, 9, kind="method"),
+ ]
+ },
+ )
+ try:
+ assert index.declared_symbol_at("app/api.py", 2).name == "OrderService"
+ finally:
+ await index.close()
+
+ async def test_a_distant_definition_is_out_of_reach(self, tmp_path: Path) -> None:
+ index = await make_repo_index(tmp_path, {"app/api.py": [_symbol("late", 60, 70)]})
+ try:
+ assert index.declared_symbol_at("app/api.py", 2) is None
+ finally:
+ await index.close()
+
+
+class TestBindSymbolIds:
+ async def test_a_line_becomes_a_symbol_id(self, tmp_path: Path) -> None:
+ index = await make_repo_index(tmp_path, {"app/api.py": [_symbol("users", 11, 14)]})
+ try:
+ contracts = [_contract(10)]
+ bind_symbol_ids(contracts, index)
+ assert contracts[0].symbol_id == "app/api.py::users"
+ assert contracts[0].symbol_name == "fastapi:GET /users", "display label is untouched"
+ finally:
+ await index.close()
+
+ async def test_an_already_bound_contract_is_left_alone(self, tmp_path: Path) -> None:
+ """The index-backed dialects bind exactly; no line rule may overrule them."""
+ index = await make_repo_index(tmp_path, {"app/api.py": [_symbol("users", 11, 14)]})
+ try:
+ contract = _contract(10)
+ contract.symbol_id = "app/api.py::chosen_by_the_dialect"
+ bind_symbol_ids([contract], index)
+ assert contract.symbol_id == "app/api.py::chosen_by_the_dialect"
+ finally:
+ await index.close()
+
+ async def test_a_contract_without_a_line_stays_unbound(self, tmp_path: Path) -> None:
+ index = await make_repo_index(tmp_path, {"app/api.py": [_symbol("users", 1, 99)]})
+ try:
+ contracts = [_contract(None)]
+ bind_symbol_ids(contracts, index)
+ assert contracts[0].symbol_id is None
+ finally:
+ await index.close()
+
+ def test_no_index_binds_nothing_and_counts_every_row_unindexed(self) -> None:
+ contracts = [_contract(10), _contract(10, role="consumer")]
+ counts = bind_symbol_ids(contracts, None)
+ assert all(c.symbol_id is None for c in contracts)
+ assert counts == {"identity_unindexed_provider": 1, "identity_unindexed_consumer": 1}
+
+ async def test_a_file_the_parser_never_saw_is_counted_separately(
+ self, tmp_path: Path
+ ) -> None:
+ """A ``.sql`` file has no symbols, so no binding rule can reach it."""
+ index = await make_repo_index(tmp_path, {"app/api.py": [_symbol("users", 11, 14)]})
+ try:
+ contracts = [_contract(10), _contract(4, path="db/schema.sql")]
+ counts = bind_symbol_ids(contracts, index)
+ assert contracts[0].symbol_id == "app/api.py::users"
+ assert contracts[1].symbol_id is None
+ assert counts == {"identity_unindexed_provider": 1}
+ finally:
+ await index.close()
+
+ async def test_a_table_declared_in_a_class_body_binds_to_the_class(
+ self, tmp_path: Path
+ ) -> None:
+ """A ``__tablename__`` is a member, so the lookahead must not apply and
+ take the first method under it."""
+ index = await make_repo_index(
+ tmp_path,
+ {
+ "app/models.py": [
+ _symbol("Order", 1, 20, path="app/models.py", kind="class"),
+ _symbol("label", 4, 6, path="app/models.py", kind="method"),
+ ]
+ },
+ )
+ try:
+ contract = _contract(2, path="app/models.py")
+ contract.contract_type = "data"
+ contract.contract_id = "data::orders"
+ bind_symbol_ids([contract], index)
+ assert contract.symbol_id == "app/models.py::Order"
+ finally:
+ await index.close()
+
+ async def test_a_module_level_call_does_not_claim_the_next_definition(
+ self, tmp_path: Path
+ ) -> None:
+ """A consumer is a call, so nothing below the line can be its symbol."""
+ index = await make_repo_index(tmp_path, {"app/api.py": [_symbol("load_orders", 6, 9)]})
+ try:
+ contract = _contract(3, role="consumer")
+ bind_symbol_ids([contract], index)
+ assert contract.symbol_id is None
+ finally:
+ await index.close()
+
+ async def test_an_indexed_file_with_nothing_at_the_line_is_not_called_unindexed(
+ self, tmp_path: Path
+ ) -> None:
+ """Reported honestly as a binding miss, not blamed on a missing parse."""
+ index = await make_repo_index(tmp_path, {"app/api.py": [_symbol("users", 60, 70)]})
+ try:
+ counts = bind_symbol_ids([_contract(2)], index)
+ assert counts == {}
+ finally:
+ await index.close()
+
+
+class TestReportedLines:
+ r"""The line a dialect reports must be the declaration's own.
+
+ Three ORM patterns open with ``^\s*`` under ``MULTILINE``, where ``^``
+ matches at a preceding blank line and ``\s*`` eats the newlines. Taking
+ the match's start rather than its group put the contract above its class.
+ """
+
+ def test_a_tablename_after_a_blank_line_reports_its_own_line(self) -> None:
+ from repowise.core.workspace.extractors.base import ScanContext
+ from repowise.core.workspace.extractors.data.orm_models import SqlAlchemyDjangoDialect
+
+ content = 'class Order(Base):\n\n __tablename__ = "orders"\n'
+ ctx = ScanContext("backend", "app/models.py", ".py", content)
+ contracts = SqlAlchemyDjangoDialect().extract(ctx)
+
+ assert [c.line for c in contracts] == [3]
+
+ def test_a_rails_model_after_blank_lines_reports_its_own_line(self) -> None:
+ from repowise.core.workspace.extractors.base import ScanContext
+ from repowise.core.workspace.extractors.data.orm_models import ActiveRecordDialect
+
+ content = "module X\nend\n\n\nclass Order < ApplicationRecord\nend\n"
+ ctx = ScanContext("backend", "app/models/order.rb", ".rb", content)
+ contracts = ActiveRecordDialect().extract(ctx)
+
+ assert [c.line for c in contracts] == [5]
+
+
+# ---------------------------------------------------------------------------
+# The artifact
+# ---------------------------------------------------------------------------
+
+
+class TestSerialization:
+ def test_identity_round_trips(self) -> None:
+ contract = _contract(10)
+ contract.symbol_id = "app/api.py::users"
+ restored = Contract.from_dict(contract.to_dict())
+ assert restored.symbol_id == "app/api.py::users"
+ assert restored.line == 10
+
+ def test_an_unbound_contract_carries_no_empty_keys(self) -> None:
+ d = _contract(None).to_dict()
+ assert "symbol_id" not in d
+ assert "line" not in d
+
+ def test_a_v1_contract_still_loads(self) -> None:
+ """The older shape: no line, no symbol id, and still a valid contract."""
+ v1 = {
+ "repo": "backend",
+ "contract_id": "http::GET::/users",
+ "contract_type": "http",
+ "role": "provider",
+ "file_path": "app/api.py",
+ "symbol_name": "fastapi:GET /users",
+ "confidence": 0.85,
+ }
+ contract = Contract.from_dict(v1)
+ assert contract.symbol_id is None
+ assert contract.line is None
+ assert contract.contract_id == "http::GET::/users"
+
+ def test_a_fresh_store_declares_the_identity_version(self) -> None:
+ assert ContractStore().version == CONTRACTS_VERSION
+ assert ContractStore.from_dict({}).version == 1, "an unversioned artifact is v1"
+
+
+class TestDiagnostics:
+ def test_the_ratio_is_reported_against_both_denominators(self) -> None:
+ bound = _contract(10)
+ bound.symbol_id = "app/api.py::users"
+ contracts = [bound, _contract(4, path="db/schema.sql"), _contract(2)]
+ stats = {"backend": {"identity_unindexed_provider": 1}}
+
+ identity = build_diagnostics(contracts, [], stats).to_dict()["symbol_identity"]
+
+ assert identity["provider"]["total"] == 3
+ assert identity["provider"]["bound"] == 1
+ assert identity["provider"]["unindexed_file"] == 1
+ assert identity["provider"]["bound_ratio"] == pytest.approx(1 / 3)
+ assert identity["provider"]["bound_ratio_indexed"] == pytest.approx(1 / 2)
+ assert identity["consumer"]["bound_ratio"] is None, "0/0 is not 100%"
+
+
+# ---------------------------------------------------------------------------
+# End to end
+# ---------------------------------------------------------------------------
+
+
+ROUTE_SOURCE = (
+ "from fastapi import APIRouter\n"
+ "\n"
+ "router = APIRouter()\n"
+ "\n"
+ '@router.get("/users")\n'
+ "async def list_users():\n"
+ " return []\n"
+)
+
+
+def two_repos(tmp_path: Path) -> WorkspaceConfig:
+ for alias in ("alpha", "beta"):
+ repo = tmp_path / alias
+ (repo / "app").mkdir(parents=True)
+ (repo / "app" / "api.py").write_text(ROUTE_SOURCE, encoding="utf-8")
+ (repo / ".repowise").mkdir()
+ return WorkspaceConfig(
+ repos=[RepoEntry(path="alpha", alias="alpha"), RepoEntry(path="beta", alias="beta")],
+ contracts=ContractConfig(),
+ )
+
+
+async def test_a_regex_extracted_route_binds_to_its_handler(
+ two_repos: WorkspaceConfig, tmp_path: Path, monkeypatch: pytest.MonkeyPatch
+) -> None:
+ """The whole path, from a dialect's match offset to the persisted id."""
+ from repowise.core.workspace import contracts as contracts_mod
+ from repowise.core.workspace.repo_index import WorkspaceIndex
+
+ monkeypatch.setattr(contracts_mod, "save_contract_store", lambda store, root: root)
+ index = await make_repo_index(
+ tmp_path / "alpha", {"app/api.py": [_symbol("list_users", 6, 7)]}, alias="alpha"
+ )
+ try:
+ store = await run_contract_extraction(
+ two_repos, tmp_path, [], workspace_index=WorkspaceIndex({"alpha": index})
+ )
+ finally:
+ await index.close()
+
+ alpha = [c for c in store.contracts if c.repo == "alpha" and c.role == "provider"]
+ assert alpha
+ assert all(c.symbol_id == "app/api.py::list_users" for c in alpha)
+ beta = [c for c in store.contracts if c.repo == "beta"]
+ assert beta
+ assert all(c.symbol_id is None for c in beta), "no index, no identity, still a row"
+
+
+async def test_a_v1_store_is_never_carried_forward(
+ two_repos: WorkspaceConfig, tmp_path: Path, monkeypatch: pytest.MonkeyPatch
+) -> None:
+ """Its rows have no identity and re-reading the source is the only fix."""
+ from repowise.core.workspace import contracts as contracts_mod
+ from repowise.core.workspace.extractors import base
+
+ monkeypatch.setattr(contracts_mod, "save_contract_store", lambda store, root: root)
+ walked: list[str] = []
+ original = base.iter_source_files
+
+ def recording(repo_path, wanted, exclude=None):
+ walked.append(Path(repo_path).name)
+ return original(repo_path, wanted, exclude)
+
+ monkeypatch.setattr(base, "iter_source_files", recording)
+
+ first = await run_contract_extraction(two_repos, tmp_path, [])
+ first.version = 1
+ walked.clear()
+
+ await run_contract_extraction(two_repos, tmp_path, [], None, first)
+
+ assert sorted(walked) == ["alpha", "beta"]
diff --git a/tests/unit/workspace/test_system_graph.py b/tests/unit/workspace/test_system_graph.py
index 9141e87e5..edfc592d8 100644
--- a/tests/unit/workspace/test_system_graph.py
+++ b/tests/unit/workspace/test_system_graph.py
@@ -311,6 +311,7 @@ def test_system_graph_json_shape_is_locked():
"consumers_by_layer",
"http_consumers_unresolved",
"http_consumer_coverage",
+ "symbol_identity",
}