codegraph-engine 2.3.0__py3-none-any.whl โ 3.0.1__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- codegraph/__init__.py +1 -1
- codegraph/__main__.py +7 -0
- codegraph/agent_brain.py +253 -2
- codegraph/agent_capabilities.py +1348 -67
- codegraph/agent_rules.py +2 -1
- codegraph/api_drift.py +455 -0
- codegraph/architecture_rules.py +123 -0
- codegraph/async_queue/__init__.py +50 -0
- codegraph/async_queue/detectors/__init__.py +48 -0
- codegraph/async_queue/detectors/celery.py +160 -0
- codegraph/async_queue/detectors/kafka.py +168 -0
- codegraph/async_queue/detectors/rabbitmq.py +116 -0
- codegraph/async_queue/detectors/redis.py +130 -0
- codegraph/async_queue/detectors/sqs.py +135 -0
- codegraph/async_queue/models.py +161 -0
- codegraph/async_queue/otel_models.py +74 -0
- codegraph/async_queue/otel_parser.py +337 -0
- codegraph/async_queue/reconciler.py +427 -0
- codegraph/async_queue/scanner.py +291 -0
- codegraph/async_queue/traversal.py +447 -0
- codegraph/change_impact.py +320 -110
- codegraph/changelog.py +170 -0
- codegraph/cli.py +1605 -62
- codegraph/cli_output.py +17 -2
- codegraph/context.py +135 -24
- codegraph/context_freshness.py +2 -0
- codegraph/database/__init__.py +14 -0
- codegraph/database/extractor.py +712 -17
- codegraph/database/interrogation.py +388 -2
- codegraph/database/schema_drift.py +862 -0
- codegraph/dataflow_propagation.py +118 -0
- codegraph/dead_code.py +387 -0
- codegraph/epistemic.py +15 -0
- codegraph/evidence_contract.py +34 -3
- codegraph/federation.py +245 -0
- codegraph/frameworks.py +234 -12
- codegraph/frameworks_nestjs.py +502 -0
- codegraph/frameworks_nextjs.py +337 -0
- codegraph/frameworks_react.py +457 -0
- codegraph/freshness.py +2 -0
- codegraph/frontend_html.py +109 -0
- codegraph/git.py +2 -0
- codegraph/git_state.py +2 -0
- codegraph/graph/traversal.py +150 -7
- codegraph/graph_csr.py +218 -0
- codegraph/incremental.py +150 -0
- codegraph/indexing/__init__.py +7 -1
- codegraph/indexing/classifier.py +5 -0
- codegraph/indexing/indexer.py +292 -38
- codegraph/indexing/languages/__init__.py +1 -0
- codegraph/indexing/languages/cpp.py +457 -0
- codegraph/indexing/languages/go.py +459 -0
- codegraph/indexing/languages/jvm.py +583 -0
- codegraph/indexing/languages/rust.py +507 -0
- codegraph/indexing/models.py +15 -1
- codegraph/indexing/parser.py +385 -9
- codegraph/indexing/pool.py +136 -0
- codegraph/indexing/protocols.py +220 -0
- codegraph/indexing/scanner.py +23 -0
- codegraph/indexing/templates.py +418 -0
- codegraph/interrogation.py +207 -39
- codegraph/lsp.py +318 -0
- codegraph/lsp_bridge/__init__.py +16 -0
- codegraph/lsp_bridge/bridge.py +102 -0
- codegraph/lsp_bridge/client.py +207 -0
- codegraph/lsp_bridge/models.py +71 -0
- codegraph/mcp/server.py +502 -29
- codegraph/memory/trajectory.py +282 -0
- codegraph/models.py +7 -0
- codegraph/monorepo.py +3 -0
- codegraph/multi_agent/__init__.py +13 -0
- codegraph/multi_agent/coordinator.py +229 -0
- codegraph/multi_agent/models.py +76 -0
- codegraph/optimizer.py +42 -5
- codegraph/polymorphism.py +95 -0
- codegraph/pr_review.py +296 -0
- codegraph/process_lifecycle.py +50 -12
- codegraph/reachability.py +198 -0
- codegraph/reachability_index.py +74 -0
- codegraph/refactor/__init__.py +48 -0
- codegraph/refactor/models.py +114 -0
- codegraph/refactor/patcher.py +348 -0
- codegraph/refactor/renamer.py +844 -0
- codegraph/refactor/transactions.py +414 -0
- codegraph/repair.py +233 -0
- codegraph/resolver.py +1357 -161
- codegraph/retrieval_policy.py +22 -1
- codegraph/roblox.py +2174 -0
- codegraph/route_composer.py +7 -0
- codegraph/route_topology.py +172 -0
- codegraph/runtime/__init__.py +25 -0
- codegraph/runtime/collector.py +155 -0
- codegraph/runtime/db_watcher.py +191 -0
- codegraph/runtime/ingestor.py +30 -10
- codegraph/runtime/otel_receiver.py +208 -0
- codegraph/runtime/port_inspector.py +304 -0
- codegraph/runtime/runner.py +155 -6
- codegraph/runtime/trace_linker.py +175 -0
- codegraph/sandbox.py +215 -0
- codegraph/search/hybrid.py +26 -0
- codegraph/security/supply_chain.py +188 -0
- codegraph/security/taint.py +421 -0
- codegraph/semantic_decorators.py +34 -3
- codegraph/structural_diff.py +2 -0
- codegraph/styles.py +144 -0
- codegraph/symbol_history.py +2 -0
- codegraph/test_gen.py +437 -0
- codegraph/tsconfig.py +236 -0
- codegraph/ui/__init__.py +6 -0
- codegraph/ui/server.py +876 -0
- codegraph/watcher.py +323 -0
- {codegraph_engine-2.3.0.dist-info โ codegraph_engine-3.0.1.dist-info}/METADATA +365 -67
- codegraph_engine-3.0.1.dist-info/RECORD +164 -0
- {codegraph_engine-2.3.0.dist-info โ codegraph_engine-3.0.1.dist-info}/WHEEL +1 -1
- codegraph_engine-2.3.0.dist-info/RECORD +0 -94
- {codegraph_engine-2.3.0.dist-info โ codegraph_engine-3.0.1.dist-info}/entry_points.txt +0 -0
- {codegraph_engine-2.3.0.dist-info โ codegraph_engine-3.0.1.dist-info}/licenses/LICENSE +0 -0
- {codegraph_engine-2.3.0.dist-info โ codegraph_engine-3.0.1.dist-info}/top_level.txt +0 -0
codegraph/optimizer.py
CHANGED
|
@@ -41,7 +41,7 @@ _CONFIDENCE_RANK = {
|
|
|
41
41
|
|
|
42
42
|
# Hard upper safety limits enforced even when omitted or exceeded
|
|
43
43
|
HARD_MAX_TOKENS = 20_000
|
|
44
|
-
HARD_MAX_FILES =
|
|
44
|
+
HARD_MAX_FILES = 40
|
|
45
45
|
HARD_MAX_LINES = 1_500
|
|
46
46
|
HARD_MAX_SYMBOLS = 50
|
|
47
47
|
HARD_MAX_RELATIONSHIPS = 60
|
|
@@ -54,7 +54,7 @@ class ContextBudgetSpec:
|
|
|
54
54
|
max_tokens: int = 20_000
|
|
55
55
|
max_symbols: int = 25
|
|
56
56
|
max_relationships: int = 30
|
|
57
|
-
max_files: int =
|
|
57
|
+
max_files: int = 25
|
|
58
58
|
max_lines: int = 500
|
|
59
59
|
max_depth: int = 3
|
|
60
60
|
max_package_depth: int = 2
|
|
@@ -94,7 +94,7 @@ class ContextBudget:
|
|
|
94
94
|
tool_calls: int = 1
|
|
95
95
|
max_symbols: int = 25
|
|
96
96
|
max_relationships: int = 30
|
|
97
|
-
max_files: int =
|
|
97
|
+
max_files: int = 25
|
|
98
98
|
max_lines: int = 500
|
|
99
99
|
max_depth: int = 3
|
|
100
100
|
max_package_depth: int = 2
|
|
@@ -374,6 +374,17 @@ def compute_item_utility(
|
|
|
374
374
|
dim_unseen = selected_dimensions is not None and dim not in selected_dimensions and dim != "GENERAL"
|
|
375
375
|
coverage_factor = 1.3 if (layer_unseen or dim_unseen) else 1.0
|
|
376
376
|
|
|
377
|
+
# Priority boost for task-critical items over tangential ones
|
|
378
|
+
priority_dim_boost = 0.0
|
|
379
|
+
if dim == "TARGET" or item.relevance_score >= 0.80:
|
|
380
|
+
priority_dim_boost = 0.20
|
|
381
|
+
elif dim in ("ROUTE", "CALLER"):
|
|
382
|
+
priority_dim_boost = 0.12
|
|
383
|
+
elif dim == "CALLEE":
|
|
384
|
+
priority_dim_boost = 0.08
|
|
385
|
+
elif dim in ("TEST", "GIT", "PACKAGE", "GENERAL") and item.relevance_score < 0.50:
|
|
386
|
+
priority_dim_boost = -0.08
|
|
387
|
+
|
|
377
388
|
# Redundancy penalty: heavily discount items sharing duplicate keys with already selected items
|
|
378
389
|
redundancy_penalty = 0.5 if item.duplicate_key in selected_duplicate_keys else 0.0
|
|
379
390
|
|
|
@@ -384,6 +395,7 @@ def compute_item_utility(
|
|
|
384
395
|
+ (fresh_factor * 0.15)
|
|
385
396
|
+ (epistemic_factor * 0.10)
|
|
386
397
|
+ pkg_adj
|
|
398
|
+
+ priority_dim_boost
|
|
387
399
|
) * coverage_factor
|
|
388
400
|
|
|
389
401
|
final_score = max(0.01, utility - redundancy_penalty - conf_penalty - art_penalty - dist_penalty)
|
|
@@ -572,7 +584,7 @@ def optimize_context_budget(
|
|
|
572
584
|
"""Deterministically select the highest-utility, coverage-preserving subset of candidates."""
|
|
573
585
|
raw_max_symbols = budget_spec.max_symbols if budget_spec else 25
|
|
574
586
|
raw_max_relationships = budget_spec.max_relationships if budget_spec else 30
|
|
575
|
-
raw_max_files = budget_spec.max_files if budget_spec else
|
|
587
|
+
raw_max_files = budget_spec.max_files if budget_spec else 25
|
|
576
588
|
raw_max_lines = budget_spec.max_lines if budget_spec else 500
|
|
577
589
|
max_depth = budget_spec.max_depth if budget_spec else 3
|
|
578
590
|
max_package_depth = budget_spec.max_package_depth if budget_spec else 2
|
|
@@ -767,7 +779,32 @@ def optimize_context_budget(
|
|
|
767
779
|
)
|
|
768
780
|
break
|
|
769
781
|
|
|
770
|
-
# Pass 1b:
|
|
782
|
+
# Pass 1b: Task-Critical Candidates Priority pass โ ensure all direct targets,
|
|
783
|
+
# routes, and direct callers/callees are admitted into the file budget before
|
|
784
|
+
# tangential layer coverage files consume file slots.
|
|
785
|
+
for item in sorted_candidates:
|
|
786
|
+
if item in selected or _is_excluded(item, task_spec):
|
|
787
|
+
continue
|
|
788
|
+
if item.duplicate_key in selected_duplicate_keys:
|
|
789
|
+
continue
|
|
790
|
+
dim = item.infer_dimension()
|
|
791
|
+
is_task_critical = (
|
|
792
|
+
dim in ("TARGET", "ROUTE", "CALLER")
|
|
793
|
+
or item.relevance_score >= 0.70
|
|
794
|
+
or (dim == "CALLEE" and item.relevance_score >= 0.50)
|
|
795
|
+
)
|
|
796
|
+
if is_task_critical:
|
|
797
|
+
if _can_fit(item) or (dim == "TARGET" and len(selected) == 0 and _trim_to_fit(item)):
|
|
798
|
+
_accept(
|
|
799
|
+
item,
|
|
800
|
+
is_coverage_fill=False,
|
|
801
|
+
why=[
|
|
802
|
+
f"Task-critical priority candidate: dimension '{dim}'",
|
|
803
|
+
f"Relevance: {item.relevance_score:.2f}, Evidence: {item.evidence_quality:.2f}",
|
|
804
|
+
],
|
|
805
|
+
)
|
|
806
|
+
|
|
807
|
+
# Pass 1c: Architectural Layer Coverage pass โ ensure distinct layers
|
|
771
808
|
# (e.g. ENTRYPOINT, SERVICE, DATA, TEST, GIT) get represented if candidates exist
|
|
772
809
|
for item in sorted_candidates:
|
|
773
810
|
if item in selected or _is_excluded(item, task_spec):
|
|
@@ -0,0 +1,95 @@
|
|
|
1
|
+
"""Polymorphic Trait & Interface Dynamic Dispatch Resolver for CodeGraph.
|
|
2
|
+
|
|
3
|
+
Resolves abstract interface and trait invocations to concrete implementing classes:
|
|
4
|
+
- Resolves Java/Kotlin interface methods (e.g. PaymentGateway.charge() -> StripeGateway.charge(), PayPalGateway.charge())
|
|
5
|
+
- Resolves Rust trait implementations (e.g. Handler::call -> AuthHandler::call)
|
|
6
|
+
- Resolves Go interface method dispatch via structural matching
|
|
7
|
+
- Resolves C++ pure virtual method overrides
|
|
8
|
+
"""
|
|
9
|
+
from __future__ import annotations
|
|
10
|
+
|
|
11
|
+
import sqlite3
|
|
12
|
+
from dataclasses import asdict, dataclass, field
|
|
13
|
+
from typing import Any
|
|
14
|
+
|
|
15
|
+
|
|
16
|
+
@dataclass(frozen=True)
|
|
17
|
+
class ConcreteMethodTarget:
|
|
18
|
+
"""A concrete class or struct implementing an abstract interface method."""
|
|
19
|
+
|
|
20
|
+
implementing_class: str
|
|
21
|
+
method_canonical_id: str
|
|
22
|
+
file_path: str
|
|
23
|
+
line: int
|
|
24
|
+
relationship: str = "IMPLEMENTS"
|
|
25
|
+
evidence_class: str = "AST_VERIFIED"
|
|
26
|
+
|
|
27
|
+
def as_dict(self) -> dict[str, Any]:
|
|
28
|
+
return asdict(self)
|
|
29
|
+
|
|
30
|
+
|
|
31
|
+
@dataclass
|
|
32
|
+
class PolymorphicResolutionReport:
|
|
33
|
+
"""Report resolving an abstract interface call to all concrete implementers."""
|
|
34
|
+
|
|
35
|
+
interface_symbol: str
|
|
36
|
+
method_name: str
|
|
37
|
+
total_implementations: int
|
|
38
|
+
implementations: list[ConcreteMethodTarget] = field(default_factory=list)
|
|
39
|
+
|
|
40
|
+
def as_dict(self) -> dict[str, Any]:
|
|
41
|
+
return {
|
|
42
|
+
"interface_symbol": self.interface_symbol,
|
|
43
|
+
"method_name": self.method_name,
|
|
44
|
+
"total_implementations": self.total_implementations,
|
|
45
|
+
"implementations": [i.as_dict() for i in self.implementations],
|
|
46
|
+
}
|
|
47
|
+
|
|
48
|
+
|
|
49
|
+
def resolve_polymorphic_callees(
|
|
50
|
+
con: sqlite3.Connection,
|
|
51
|
+
interface_symbol: str,
|
|
52
|
+
method_name: str,
|
|
53
|
+
) -> PolymorphicResolutionReport:
|
|
54
|
+
"""Resolve an interface/trait method to all concrete implementations in the workspace."""
|
|
55
|
+
cur = con.cursor()
|
|
56
|
+
|
|
57
|
+
# 1. Find all classes that inherit or implement interface_symbol
|
|
58
|
+
cur.execute("""
|
|
59
|
+
SELECT source_symbol, source_canonical_id, source_file, line
|
|
60
|
+
FROM inheritance
|
|
61
|
+
WHERE base_name = ? OR base_name LIKE ?;
|
|
62
|
+
""", (interface_symbol, f"%{interface_symbol}"))
|
|
63
|
+
implementing_types = cur.fetchall()
|
|
64
|
+
|
|
65
|
+
targets: list[ConcreteMethodTarget] = []
|
|
66
|
+
|
|
67
|
+
for sub_name, _sub_cid, sub_file, _ in implementing_types:
|
|
68
|
+
# 2. Look for the method on the implementing type
|
|
69
|
+
expected_method_qname = f"{sub_name}.{method_name}"
|
|
70
|
+
cur.execute("""
|
|
71
|
+
SELECT canonical_id, path, start_line
|
|
72
|
+
FROM symbols
|
|
73
|
+
WHERE (name = ? OR qualified_name = ?) AND (path = ? OR 1=1)
|
|
74
|
+
LIMIT 1;
|
|
75
|
+
""", (method_name, expected_method_qname, sub_file))
|
|
76
|
+
m_row = cur.fetchone()
|
|
77
|
+
|
|
78
|
+
if m_row:
|
|
79
|
+
targets.append(
|
|
80
|
+
ConcreteMethodTarget(
|
|
81
|
+
implementing_class=sub_name,
|
|
82
|
+
method_canonical_id=str(m_row[0]),
|
|
83
|
+
file_path=str(m_row[1]),
|
|
84
|
+
line=int(m_row[2]),
|
|
85
|
+
relationship="IMPLEMENTS",
|
|
86
|
+
evidence_class="AST_VERIFIED",
|
|
87
|
+
)
|
|
88
|
+
)
|
|
89
|
+
|
|
90
|
+
return PolymorphicResolutionReport(
|
|
91
|
+
interface_symbol=interface_symbol,
|
|
92
|
+
method_name=method_name,
|
|
93
|
+
total_implementations=len(targets),
|
|
94
|
+
implementations=targets,
|
|
95
|
+
)
|
codegraph/pr_review.py
ADDED
|
@@ -0,0 +1,296 @@
|
|
|
1
|
+
"""Native Pull Request Review & Semantic Blast-Radius Auditor for CodeGraph.
|
|
2
|
+
|
|
3
|
+
Performs deep structural AST diffs, caller contract verification, affected HTTP route
|
|
4
|
+
detection, database schema impact analysis, and test coverage gap identification to generate
|
|
5
|
+
an actionable, high-density PR review report for CI/CD pipelines (GitHub Actions, GitLab CI).
|
|
6
|
+
"""
|
|
7
|
+
from __future__ import annotations
|
|
8
|
+
|
|
9
|
+
import sqlite3
|
|
10
|
+
from dataclasses import dataclass, field
|
|
11
|
+
from enum import StrEnum
|
|
12
|
+
from pathlib import Path
|
|
13
|
+
from typing import Any
|
|
14
|
+
|
|
15
|
+
from codegraph.change_impact import get_deep_change_impact
|
|
16
|
+
from codegraph.semantic_conflicts import detect_semantic_conflicts
|
|
17
|
+
from codegraph.structural_diff import compare_branches
|
|
18
|
+
|
|
19
|
+
|
|
20
|
+
class ReviewSeverity(StrEnum):
|
|
21
|
+
CRITICAL = "CRITICAL"
|
|
22
|
+
HIGH = "HIGH"
|
|
23
|
+
MEDIUM = "MEDIUM"
|
|
24
|
+
LOW = "LOW"
|
|
25
|
+
INFO = "INFO"
|
|
26
|
+
|
|
27
|
+
|
|
28
|
+
@dataclass
|
|
29
|
+
class ReviewFinding:
|
|
30
|
+
severity: ReviewSeverity
|
|
31
|
+
category: str
|
|
32
|
+
title: str
|
|
33
|
+
description: str
|
|
34
|
+
file_path: str
|
|
35
|
+
line: int | None = None
|
|
36
|
+
recommendation: str = ""
|
|
37
|
+
|
|
38
|
+
def as_dict(self) -> dict[str, Any]:
|
|
39
|
+
return {
|
|
40
|
+
"severity": self.severity.value,
|
|
41
|
+
"category": self.category,
|
|
42
|
+
"title": self.title,
|
|
43
|
+
"description": self.description,
|
|
44
|
+
"file_path": self.file_path,
|
|
45
|
+
"line": self.line,
|
|
46
|
+
"recommendation": self.recommendation,
|
|
47
|
+
}
|
|
48
|
+
|
|
49
|
+
|
|
50
|
+
@dataclass
|
|
51
|
+
class PRReviewReport:
|
|
52
|
+
base_ref: str
|
|
53
|
+
head_ref: str
|
|
54
|
+
verdict: str # APPROVE, COMMENT, REQUEST_CHANGES
|
|
55
|
+
summary: str
|
|
56
|
+
findings: list[ReviewFinding] = field(default_factory=list)
|
|
57
|
+
files_changed: list[str] = field(default_factory=list)
|
|
58
|
+
symbols_added: int = 0
|
|
59
|
+
symbols_modified: int = 0
|
|
60
|
+
symbols_removed: int = 0
|
|
61
|
+
affected_routes: list[dict[str, Any]] = field(default_factory=list)
|
|
62
|
+
affected_tests: list[str] = field(default_factory=list)
|
|
63
|
+
suggested_test_command: str = ""
|
|
64
|
+
markdown_report: str = ""
|
|
65
|
+
|
|
66
|
+
def as_dict(self) -> dict[str, Any]:
|
|
67
|
+
return {
|
|
68
|
+
"base_ref": self.base_ref,
|
|
69
|
+
"head_ref": self.head_ref,
|
|
70
|
+
"verdict": self.verdict,
|
|
71
|
+
"summary": self.summary,
|
|
72
|
+
"findings": [f.as_dict() for f in self.findings],
|
|
73
|
+
"files_changed": self.files_changed,
|
|
74
|
+
"symbols_added": self.symbols_added,
|
|
75
|
+
"symbols_modified": self.symbols_modified,
|
|
76
|
+
"symbols_removed": self.symbols_removed,
|
|
77
|
+
"affected_routes": self.affected_routes,
|
|
78
|
+
"affected_tests": self.affected_tests,
|
|
79
|
+
"suggested_test_command": self.suggested_test_command,
|
|
80
|
+
"markdown_report": self.markdown_report,
|
|
81
|
+
}
|
|
82
|
+
|
|
83
|
+
|
|
84
|
+
def generate_pr_review(
|
|
85
|
+
repository: Path,
|
|
86
|
+
base_ref: str = "main",
|
|
87
|
+
head_ref: str = "HEAD",
|
|
88
|
+
con: sqlite3.Connection | None = None,
|
|
89
|
+
) -> PRReviewReport:
|
|
90
|
+
"""Generate a comprehensive semantic PR review evaluating structural diffs, contract breaks, and tests."""
|
|
91
|
+
repo = repository.resolve()
|
|
92
|
+
|
|
93
|
+
# 1. Structural AST Diff
|
|
94
|
+
diff_report = compare_branches(repo, base_branch=base_ref, head_branch=head_ref, con=con)
|
|
95
|
+
|
|
96
|
+
# 2. Semantic Conflict & Contract Break Analysis
|
|
97
|
+
conflicts_report = detect_semantic_conflicts(repo, base_branch=base_ref, head_branch=head_ref, con=con)
|
|
98
|
+
|
|
99
|
+
# 3. Deep Change Impact Analysis
|
|
100
|
+
files_changed = sorted(
|
|
101
|
+
set(diff_report.added_files + diff_report.modified_files + diff_report.deleted_files)
|
|
102
|
+
)
|
|
103
|
+
impact_report = None
|
|
104
|
+
if files_changed and con is not None:
|
|
105
|
+
try:
|
|
106
|
+
impact_report = get_deep_change_impact(repo, con, base=base_ref, head=head_ref)
|
|
107
|
+
except Exception:
|
|
108
|
+
impact_report = None
|
|
109
|
+
|
|
110
|
+
# 4. Synthesize Findings
|
|
111
|
+
findings: list[ReviewFinding] = []
|
|
112
|
+
|
|
113
|
+
# Map semantic contract breaks
|
|
114
|
+
for conf in conflicts_report.conflicts:
|
|
115
|
+
sev = ReviewSeverity.CRITICAL if conf.severity == "CRITICAL" else ReviewSeverity.HIGH
|
|
116
|
+
findings.append(
|
|
117
|
+
ReviewFinding(
|
|
118
|
+
severity=sev,
|
|
119
|
+
category="CONTRACT_BREAK",
|
|
120
|
+
title=f"Semantic Contract Break: {conf.symbol}",
|
|
121
|
+
description=conf.description,
|
|
122
|
+
file_path=conf.head_file or conf.base_file,
|
|
123
|
+
line=conf.line,
|
|
124
|
+
recommendation="Update consuming call sites or maintain backward-compatible argument signatures.",
|
|
125
|
+
)
|
|
126
|
+
)
|
|
127
|
+
|
|
128
|
+
# Map removed symbols without replacement
|
|
129
|
+
for sym in diff_report.removed_symbols:
|
|
130
|
+
findings.append(
|
|
131
|
+
ReviewFinding(
|
|
132
|
+
severity=ReviewSeverity.HIGH,
|
|
133
|
+
category="DELETED_SYMBOL",
|
|
134
|
+
title=f"Deleted Symbol: {sym.name}",
|
|
135
|
+
description=f"{sym.kind} '{sym.name}' was removed from {sym.path}.",
|
|
136
|
+
file_path=sym.path,
|
|
137
|
+
line=sym.new_line or sym.old_line,
|
|
138
|
+
recommendation="Verify that no external packages or unindexed scripts rely on this symbol.",
|
|
139
|
+
)
|
|
140
|
+
)
|
|
141
|
+
|
|
142
|
+
# Map modified routes & missing test coverage
|
|
143
|
+
affected_routes_list: list[dict[str, Any]] = []
|
|
144
|
+
for r in diff_report.changed_routes:
|
|
145
|
+
affected_routes_list.append(r.as_dict())
|
|
146
|
+
|
|
147
|
+
if impact_report:
|
|
148
|
+
for r_item in impact_report.affected_routes:
|
|
149
|
+
r_dict = r_item.as_dict()
|
|
150
|
+
if r_dict not in affected_routes_list:
|
|
151
|
+
affected_routes_list.append(r_dict)
|
|
152
|
+
|
|
153
|
+
# Test coverage gap analysis: Check if modified routes have tests
|
|
154
|
+
covered_test_paths = set(diff_report.affected_tests)
|
|
155
|
+
if impact_report:
|
|
156
|
+
covered_test_paths.update(t.file for t in impact_report.affected_tests)
|
|
157
|
+
|
|
158
|
+
if affected_routes_list and not covered_test_paths:
|
|
159
|
+
findings.append(
|
|
160
|
+
ReviewFinding(
|
|
161
|
+
severity=ReviewSeverity.HIGH,
|
|
162
|
+
category="TEST_GAP",
|
|
163
|
+
title="Uncovered Route Modifications",
|
|
164
|
+
description=f"{len(affected_routes_list)} HTTP route(s) were added or modified without any associated test coverage.",
|
|
165
|
+
file_path=str(affected_routes_list[0].get("file_path") or "routes"),
|
|
166
|
+
recommendation="Add integration tests exercising the modified HTTP endpoints.",
|
|
167
|
+
)
|
|
168
|
+
)
|
|
169
|
+
|
|
170
|
+
# Verdict calculation
|
|
171
|
+
has_critical = any(f.severity == ReviewSeverity.CRITICAL for f in findings)
|
|
172
|
+
has_high = any(f.severity == ReviewSeverity.HIGH for f in findings)
|
|
173
|
+
|
|
174
|
+
if has_critical:
|
|
175
|
+
verdict = "REQUEST_CHANGES"
|
|
176
|
+
summary = f"Blocked: {len([f for f in findings if f.severity == ReviewSeverity.CRITICAL])} critical contract break(s) detected."
|
|
177
|
+
elif has_high:
|
|
178
|
+
verdict = "COMMENT"
|
|
179
|
+
summary = f"Caution: {len([f for f in findings if f.severity == ReviewSeverity.HIGH])} high-priority finding(s) require verification."
|
|
180
|
+
else:
|
|
181
|
+
verdict = "APPROVE"
|
|
182
|
+
summary = "Clean: No semantic contract breaks or uncovered route regressions detected."
|
|
183
|
+
|
|
184
|
+
# Build suggested test command
|
|
185
|
+
suggested_test_cmd = ""
|
|
186
|
+
if impact_report and impact_report.recommended_test_command:
|
|
187
|
+
suggested_test_cmd = impact_report.recommended_test_command
|
|
188
|
+
elif covered_test_paths:
|
|
189
|
+
suggested_test_cmd = f"pytest {' '.join(sorted(covered_test_paths)[:10])}"
|
|
190
|
+
|
|
191
|
+
# Build High-Density GitHub PR Markdown
|
|
192
|
+
markdown = _format_pr_markdown(
|
|
193
|
+
base_ref=base_ref,
|
|
194
|
+
head_ref=head_ref,
|
|
195
|
+
verdict=verdict,
|
|
196
|
+
summary=summary,
|
|
197
|
+
findings=findings,
|
|
198
|
+
files_changed=files_changed,
|
|
199
|
+
symbols_added=len(diff_report.added_symbols),
|
|
200
|
+
symbols_modified=len(diff_report.changed_symbols),
|
|
201
|
+
symbols_removed=len(diff_report.removed_symbols),
|
|
202
|
+
affected_routes=affected_routes_list,
|
|
203
|
+
affected_tests=sorted(covered_test_paths),
|
|
204
|
+
suggested_test_command=suggested_test_cmd,
|
|
205
|
+
)
|
|
206
|
+
|
|
207
|
+
return PRReviewReport(
|
|
208
|
+
base_ref=base_ref,
|
|
209
|
+
head_ref=head_ref,
|
|
210
|
+
verdict=verdict,
|
|
211
|
+
summary=summary,
|
|
212
|
+
findings=findings,
|
|
213
|
+
files_changed=files_changed,
|
|
214
|
+
symbols_added=len(diff_report.added_symbols),
|
|
215
|
+
symbols_modified=len(diff_report.changed_symbols),
|
|
216
|
+
symbols_removed=len(diff_report.removed_symbols),
|
|
217
|
+
affected_routes=affected_routes_list,
|
|
218
|
+
affected_tests=sorted(covered_test_paths),
|
|
219
|
+
suggested_test_command=suggested_test_cmd,
|
|
220
|
+
markdown_report=markdown,
|
|
221
|
+
)
|
|
222
|
+
|
|
223
|
+
|
|
224
|
+
def _format_pr_markdown(
|
|
225
|
+
base_ref: str,
|
|
226
|
+
head_ref: str,
|
|
227
|
+
verdict: str,
|
|
228
|
+
summary: str,
|
|
229
|
+
findings: list[ReviewFinding],
|
|
230
|
+
files_changed: list[str],
|
|
231
|
+
symbols_added: int,
|
|
232
|
+
symbols_modified: int,
|
|
233
|
+
symbols_removed: int,
|
|
234
|
+
affected_routes: list[dict[str, Any]],
|
|
235
|
+
affected_tests: list[str],
|
|
236
|
+
suggested_test_command: str,
|
|
237
|
+
) -> str:
|
|
238
|
+
"""Format markdown for GitHub PR review comment."""
|
|
239
|
+
verdict_badge = {
|
|
240
|
+
"APPROVE": "๐ข **APPROVED**",
|
|
241
|
+
"COMMENT": "๐ก **CAUTION / COMMENT**",
|
|
242
|
+
"REQUEST_CHANGES": "๐ด **CHANGES REQUESTED**",
|
|
243
|
+
}.get(verdict, verdict)
|
|
244
|
+
|
|
245
|
+
lines: list[str] = [
|
|
246
|
+
"## ๐ CodeGraph Semantic PR Review",
|
|
247
|
+
"",
|
|
248
|
+
f"**Verdict**: {verdict_badge} — {summary}",
|
|
249
|
+
f"**Revisions**: `{base_ref}` → `{head_ref}`",
|
|
250
|
+
"",
|
|
251
|
+
"### ๐ Blast-Radius & AST Metrics",
|
|
252
|
+
"",
|
|
253
|
+
"| Files Changed | Symbols Added | Symbols Modified | Symbols Removed | Affected Routes | Covered Tests |",
|
|
254
|
+
"| :---: | :---: | :---: | :---: | :---: | :---: |",
|
|
255
|
+
f"| {len(files_changed)} | {symbols_added} | {symbols_modified} | {symbols_removed} | {len(affected_routes)} | {len(affected_tests)} |",
|
|
256
|
+
"",
|
|
257
|
+
]
|
|
258
|
+
|
|
259
|
+
if findings:
|
|
260
|
+
lines.append("### โ ๏ธ Key Findings")
|
|
261
|
+
lines.append("")
|
|
262
|
+
for f in findings:
|
|
263
|
+
badge = "๐ด" if f.severity == ReviewSeverity.CRITICAL else ("๐ " if f.severity == ReviewSeverity.HIGH else "๐ต")
|
|
264
|
+
loc = f"`{f.file_path}:{f.line}`" if f.line else f"`{f.file_path}`"
|
|
265
|
+
lines.append(f"- {badge} **[{f.severity.value}] {f.title}** ({loc})")
|
|
266
|
+
lines.append(f" {f.description}")
|
|
267
|
+
if f.recommendation:
|
|
268
|
+
lines.append(f" *Recommendation*: {f.recommendation}")
|
|
269
|
+
lines.append("")
|
|
270
|
+
|
|
271
|
+
if affected_routes:
|
|
272
|
+
lines.append("<details><summary><b>๐ Affected HTTP Routes (" + str(len(affected_routes)) + ")</b></summary>")
|
|
273
|
+
lines.append("")
|
|
274
|
+
lines.append("| Method | Path | Handler | File |")
|
|
275
|
+
lines.append("| :--- | :--- | :--- | :--- |")
|
|
276
|
+
for r in affected_routes[:20]:
|
|
277
|
+
m = r.get("method") or r.get("http_method") or "ANY"
|
|
278
|
+
p = r.get("path") or r.get("route_path") or "-"
|
|
279
|
+
h = r.get("handler") or r.get("handler_name") or "-"
|
|
280
|
+
f_path = r.get("file_path") or "-"
|
|
281
|
+
lines.append(f"| `{m}` | `{p}` | `{h}` | `{f_path}` |")
|
|
282
|
+
lines.append("")
|
|
283
|
+
lines.append("</details>")
|
|
284
|
+
lines.append("")
|
|
285
|
+
|
|
286
|
+
if suggested_test_command:
|
|
287
|
+
lines.append("### ๐งช Recommended Targeted Test Suite")
|
|
288
|
+
lines.append("```bash")
|
|
289
|
+
lines.append(suggested_test_command)
|
|
290
|
+
lines.append("```")
|
|
291
|
+
lines.append("")
|
|
292
|
+
|
|
293
|
+
lines.append("---")
|
|
294
|
+
lines.append("*Generated by [CodeGraph MCP](https://github.com/raghurammrsd/CODE_GRAPH_MCP) • Deterministic Repository Intelligence*")
|
|
295
|
+
|
|
296
|
+
return "\n".join(lines)
|
codegraph/process_lifecycle.py
CHANGED
|
@@ -316,6 +316,16 @@ def _win32_inspect_process(pid: int) -> LiveProcessInfo | None:
|
|
|
316
316
|
access = 0x00100000 | 0x00001000
|
|
317
317
|
handle = kernel32.OpenProcess(access, False, int(pid))
|
|
318
318
|
if not handle:
|
|
319
|
+
err = kernel32.GetLastError()
|
|
320
|
+
if err == 5: # ERROR_ACCESS_DENIED: process exists and is active
|
|
321
|
+
return LiveProcessInfo(
|
|
322
|
+
pid=pid,
|
|
323
|
+
ppid=0,
|
|
324
|
+
create_token="access_denied",
|
|
325
|
+
executable="python",
|
|
326
|
+
command_summary="python",
|
|
327
|
+
is_zombie=False,
|
|
328
|
+
)
|
|
319
329
|
return None
|
|
320
330
|
try:
|
|
321
331
|
exit_code = ctypes.c_ulong(0)
|
|
@@ -541,16 +551,38 @@ def is_parent_alive(
|
|
|
541
551
|
) -> bool:
|
|
542
552
|
"""Return True if the parent process `parent_pid` is still alive and owns this child."""
|
|
543
553
|
if parent_pid <= 1:
|
|
544
|
-
|
|
554
|
+
# PID 0 and 1 represent init / systemd / launchd / container root; always considered alive
|
|
555
|
+
return True
|
|
545
556
|
|
|
546
|
-
|
|
547
|
-
|
|
548
|
-
|
|
549
|
-
|
|
557
|
+
# On Windows, os.kill(pid, 0) is not supported and raises WinError 87.
|
|
558
|
+
# Use native kernel32 inspection via get_live_process_info.
|
|
559
|
+
if sys.platform == "win32":
|
|
560
|
+
live = get_live_process_info(parent_pid)
|
|
561
|
+
if live is None:
|
|
562
|
+
return False
|
|
563
|
+
if live.is_zombie:
|
|
564
|
+
return False
|
|
565
|
+
if parent_create_token and live.create_token and live.create_token != parent_create_token:
|
|
550
566
|
return False
|
|
567
|
+
return True
|
|
568
|
+
|
|
569
|
+
# 1. Direct probe whether parent PID exists via os.kill(pid, 0)
|
|
570
|
+
try:
|
|
571
|
+
os.kill(parent_pid, 0)
|
|
572
|
+
except ProcessLookupError:
|
|
573
|
+
# ESRCH: Process does not exist; parent is definitely dead
|
|
574
|
+
return False
|
|
575
|
+
except PermissionError:
|
|
576
|
+
# EPERM: Process exists, but caller lacks permissions to signal it (e.g. sandbox/different user)
|
|
577
|
+
return True
|
|
578
|
+
except OSError:
|
|
579
|
+
# Other OS error: safe fallback is alive
|
|
580
|
+
return True
|
|
551
581
|
|
|
552
582
|
live = get_live_process_info(parent_pid)
|
|
553
|
-
if live is None
|
|
583
|
+
if live is None:
|
|
584
|
+
return True
|
|
585
|
+
if live.is_zombie:
|
|
554
586
|
return False
|
|
555
587
|
|
|
556
588
|
if parent_create_token and live.create_token and live.create_token != parent_create_token:
|
|
@@ -964,16 +996,22 @@ class MCPLifecycleController:
|
|
|
964
996
|
|
|
965
997
|
def _watchdog_loop(self) -> None:
|
|
966
998
|
"""Low-overhead event-backed watchdog detecting parent termination or closed stdin."""
|
|
999
|
+
disable_parent_watchdog = (
|
|
1000
|
+
os.environ.get("CODEGRAPH_DISABLE_PARENT_WATCHDOG", "").strip().lower()
|
|
1001
|
+
in ("1", "true", "yes")
|
|
1002
|
+
or self.record.parent_pid <= 1
|
|
1003
|
+
)
|
|
967
1004
|
while not self._stop_event.wait(timeout=self.poll_interval_sec):
|
|
968
1005
|
if _is_stream_closed_or_eof(sys.stdin):
|
|
969
1006
|
self.shutdown("stdin_closed")
|
|
970
1007
|
return
|
|
971
|
-
if not
|
|
972
|
-
|
|
973
|
-
|
|
974
|
-
|
|
975
|
-
|
|
976
|
-
|
|
1008
|
+
if not disable_parent_watchdog:
|
|
1009
|
+
if not is_parent_alive(
|
|
1010
|
+
self.record.parent_pid,
|
|
1011
|
+
self.record.parent_create_token,
|
|
1012
|
+
):
|
|
1013
|
+
self.shutdown("parent_process_terminated")
|
|
1014
|
+
return
|
|
977
1015
|
|
|
978
1016
|
def _atexit_cleanup(self) -> None:
|
|
979
1017
|
self.shutdown("atexit")
|