xai-review 0.7.0__tar.gz → 0.10.0__tar.gz

{xai_review-0.7.0 → xai_review-0.10.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: xai-review
-Version: 0.7.0
+Version: 0.10.0
 Summary: AI-powered code review tool
 Author-email: Nikita Filonov <nikita.filonov@example.com>
 Maintainer-email: Nikita Filonov <nikita.filonov@example.com>
@@ -298,6 +298,6 @@ ai-review:
 
 ## 📂 Examples
 
-- [./docs/ci/gitlab.yaml](./docs/ci/gitlab.yaml) — ready-to-use CI snippet
+- [./docs/ci](./docs/ci) — ready-to-use CI snippets
 - [./docs/configs](./docs/configs) — sample `.yaml`, `.json`, `.env` configs
 - [./docs/prompts](./docs/prompts) — prompt templates for Python/Go (light & strict modes)

{xai_review-0.7.0 → xai_review-0.10.0}/README.md

@@ -274,6 +274,6 @@ ai-review:
 
 ## 📂 Examples
 
-- [./docs/ci/gitlab.yaml](./docs/ci/gitlab.yaml) — ready-to-use CI snippet
+- [./docs/ci](./docs/ci) — ready-to-use CI snippets
 - [./docs/configs](./docs/configs) — sample `.yaml`, `.json`, `.env` configs
 - [./docs/prompts](./docs/prompts) — prompt templates for Python/Go (light & strict modes)

{xai_review-0.7.0 → xai_review-0.10.0}/ai_review/clients/gitlab/client.py

@@ -13,7 +13,7 @@ class GitLabHTTPClient:
 
 
 def get_gitlab_http_client() -> GitLabHTTPClient:
-    logger = get_logger("GITLAB_MERGE_REQUESTS_HTTP_CLIENT")
+    logger = get_logger("GITLAB_HTTP_CLIENT")
     logger_event_hook = LoggerEventHook(logger=logger)
     retry_transport = RetryTransport(transport=AsyncHTTPTransport())
 

xai_review-0.10.0/ai_review/libs/json.py

@@ -0,0 +1,19 @@
+import re
+
+CONTROL_CHARS_RE = re.compile(r"[\x00-\x1F]")
+
+
+def sanitize_json_string(raw: str) -> str:
+    def replace(match: re.Match) -> str:
+        char = match.group()
+        match char:
+            case "\n":
+                return "\\n"
+            case "\r":
+                return "\\r"
+            case "\t":
+                return "\\t"
+            case _:
+                return f"\\u{ord(char):04x}"
+
+    return CONTROL_CHARS_RE.sub(replace, raw)

{xai_review-0.7.0 → xai_review-0.10.0}/ai_review/services/diff/renderers.py

@@ -13,7 +13,7 @@ Supported build modes:
 - ADDED_AND_REMOVED_WITH_CONTEXT added + removed + surrounding unchanged lines
 """
 from enum import Enum
-from typing import Iterable, Optional
+from typing import Iterable
 
 from ai_review.libs.diff.models import DiffFile, DiffLineType
 from ai_review.services.diff.tools import normalize_file_path, marker_for_line, read_snapshot
@@ -24,7 +24,7 @@ class MarkerType(Enum):
     REMOVED = "removed"
 
 
-def build_full_file_current(file: Optional[DiffFile], file_path: str, head_sha: str | None) -> str:
+def build_full_file_current(file: DiffFile | None, file_path: str, head_sha: str | None) -> str:
     text = read_snapshot(file_path, head_sha=head_sha)
     if text is None:
         return f"# Failed to read current snapshot for {file_path}"
@@ -33,7 +33,7 @@ def build_full_file_current(file: Optional[DiffFile], file_path: str, head_sha:
     return render_plain_numbered(text.splitlines(), added_new, marker_type=MarkerType.ADDED)
 
 
-def build_full_file_previous(file: Optional[DiffFile], file_path: str, base_sha: str | None) -> str:
+def build_full_file_previous(file: DiffFile | None, file_path: str, base_sha: str | None) -> str:
     text = read_snapshot(file_path, base_sha=base_sha)
     if text is None:
         return f"# Failed to read previous snapshot for {file_path} (base_sha missing or file absent)"
@@ -42,31 +42,31 @@ def build_full_file_previous(file: Optional[DiffFile], file_path: str, base_sha:
     return render_plain_numbered(text.splitlines(), removed_old, marker_type=MarkerType.REMOVED)
 
 
-def build_full_file_diff(file: DiffFile) -> str:
+def build_full_file_diff(file: DiffFile | None) -> str:
     return render_unified(file, include_added=True, include_removed=True, include_unchanged=True, context=0)
 
 
-def build_only_added(file: DiffFile) -> str:
+def build_only_added(file: DiffFile | None) -> str:
     return render_unified(file, include_added=True, include_removed=False, include_unchanged=False, context=0)
 
 
-def build_only_removed(file: DiffFile) -> str:
+def build_only_removed(file: DiffFile | None) -> str:
     return render_unified(file, include_added=False, include_removed=True, include_unchanged=False, context=0)
 
 
-def build_added_and_removed(file: DiffFile) -> str:
+def build_added_and_removed(file: DiffFile | None) -> str:
     return render_unified(file, include_added=True, include_removed=True, include_unchanged=False, context=0)
 
 
-def build_only_added_with_context(file: DiffFile, context: int) -> str:
+def build_only_added_with_context(file: DiffFile | None, context: int) -> str:
     return render_unified(file, include_added=True, include_removed=False, include_unchanged=True, context=context)
 
 
-def build_only_removed_with_context(file: DiffFile, context: int) -> str:
+def build_only_removed_with_context(file: DiffFile | None, context: int) -> str:
     return render_unified(file, include_added=False, include_removed=True, include_unchanged=True, context=context)
 
 
-def build_added_and_removed_with_context(file: DiffFile, context: int) -> str:
+def build_added_and_removed_with_context(file: DiffFile | None, context: int) -> str:
     return render_unified(file, include_added=True, include_removed=True, include_unchanged=True, context=context)
 
 
@@ -87,7 +87,7 @@ def render_plain_numbered(lines: Iterable[str], changed: set[int], marker_type:
 
 
 def render_unified(
-        file: DiffFile,
+        file: DiffFile | None,
         *,
         include_added: bool,
         include_removed: bool,
@@ -104,12 +104,19 @@ def render_unified(
 
     Context controls how many unchanged lines around modifications are shown.
     """
+    if file is None:
+        return "# Diff target not found"
+
+    if not file.hunks:
+        header = normalize_file_path(file.new_name or file.orig_name)
+        return f"# No matching lines for mode in {header}"
+
     lines_out: list[str] = []
 
     added_new_positions = file.added_line_numbers()
     removed_old_positions = file.removed_line_numbers()
 
-    def in_context(old_no: Optional[int], new_no: Optional[int]) -> bool:
+    def in_context(old_no: int | None, new_no: int | None) -> bool:
         """Check if an unchanged line falls within context radius."""
         if context <= 0:
             return False

{xai_review-0.7.0 → xai_review-0.10.0}/ai_review/services/prompt/adapter.py

@@ -12,7 +12,7 @@ def build_prompt_context_from_mr_info(mr: MRInfoSchema) -> PromptContextSchema:
 
         merge_request_reviewers=[reviewer.name for reviewer in mr.reviewers],
         merge_request_reviewers_usernames=[reviewer.username for reviewer in mr.reviewers],
-        merge_request_reviewer=mr.reviewers[0].name if mr.reviewers else None,
+        merge_request_reviewer=mr.reviewers[0].name if mr.reviewers else "",
 
         merge_request_assignees=[assignee.name for assignee in mr.assignees],
         merge_request_assignees_usernames=[assignee.username for assignee in mr.assignees],

xai_review-0.10.0/ai_review/services/review/inline/service.py

@@ -0,0 +1,53 @@
+import re
+
+from pydantic import ValidationError
+
+from ai_review.libs.json import sanitize_json_string
+from ai_review.libs.logger import get_logger
+from ai_review.services.review.inline.schema import InlineCommentListSchema
+
+logger = get_logger("INLINE_COMMENT_SERVICE")
+
+FIRST_JSON_ARRAY_RE = re.compile(r"\[[\s\S]*]", re.MULTILINE)
+CLEAN_JSON_BLOCK_RE = re.compile(r"```(?:json)?(.*?)```", re.DOTALL | re.IGNORECASE)
+
+
+class InlineCommentService:
+    @classmethod
+    def try_parse_model_output(cls, raw: str) -> InlineCommentListSchema | None:
+        try:
+            return InlineCommentListSchema.model_validate_json(raw)
+        except ValidationError as error:
+            logger.debug(f"Parse failed, trying sanitized JSON: {raw[:200]=}, {error=}")
+            try:
+                cleaned = sanitize_json_string(raw)
+                return InlineCommentListSchema.model_validate_json(cleaned)
+            except ValidationError as error:
+                logger.debug(f"Sanitized JSON still invalid: {raw[:200]=}, {error=}")
+                return None
+
+    @classmethod
+    def parse_model_output(cls, output: str) -> InlineCommentListSchema:
+        output = (output or "").strip()
+        if not output:
+            logger.warning("️LLM returned empty string for inline review")
+            return InlineCommentListSchema(root=[])
+
+        if match := CLEAN_JSON_BLOCK_RE.search(output):
+            output = match.group(1).strip()
+
+        if parsed := cls.try_parse_model_output(output):
+            return parsed
+
+        logger.warning("Failed to parse LLM output as JSON, trying to extract first JSON array...")
+
+        if json_array_match := FIRST_JSON_ARRAY_RE.search(output):
+            if parsed := cls.try_parse_model_output(json_array_match.group(0)):
+                logger.info("Successfully parsed JSON after extracting array from output")
+                return parsed
+            else:
+                logger.error("Extracted JSON array is still invalid after sanitization")
+        else:
+            logger.error("No JSON array found in LLM output")
+
+        return InlineCommentListSchema(root=[])

xai_review-0.10.0/ai_review/tests/suites/libs/test_json.py

@@ -0,0 +1,41 @@
+import pytest
+
+from ai_review.libs.json import sanitize_json_string
+
+
+@pytest.mark.parametrize(
+    ("actual", "expected"),
+    [
+        ("hello world", "hello world"),
+        ("line1\nline2", "line1\\nline2"),
+        ("foo\rbar", "foo\\rbar"),
+        ("a\tb", "a\\tb"),
+        ("abc\0def", "abc\\u0000def"),
+        ("x\n\ry\t\0z", "x\\n\\ry\\t\\u0000z"),
+    ],
+)
+def test_sanitize_basic_cases(actual: str, expected: str) -> None:
+    assert sanitize_json_string(actual) == expected
+
+
+def test_sanitize_multiple_control_chars() -> None:
+    raw = "A\nB\rC\tD\0E"
+    result = sanitize_json_string(raw)
+    assert result == "A\\nB\\rC\\tD\\u0000E"
+
+
+def test_sanitize_idempotent() -> None:
+    raw = "foo\nbar"
+    once = sanitize_json_string(raw)
+    twice = sanitize_json_string(once)
+    assert once == twice
+
+
+def test_sanitize_empty_string() -> None:
+    assert sanitize_json_string("") == ""
+
+
+def test_sanitize_only_control_chars() -> None:
+    raw = "\n\r\t\0"
+    result = sanitize_json_string(raw)
+    assert result == "\\n\\r\\t\\u0000"

{xai_review-0.7.0 → xai_review-0.10.0}/ai_review/tests/suites/services/diff/test_renderers.py

@@ -166,3 +166,26 @@ def test_build_added_and_removed_with_context(sample_diff_file: DiffFile) -> Non
         " 2: keep B\n"
         "+3: added me # added"
     )
+
+
+def test_build_full_file_diff_empty_file() -> None:
+    """
+    Should handle new empty file (mode=NEW, no hunks).
+    """
+    file = DiffFile(
+        header="diff --git a/LICENSE b/LICENSE",
+        mode=FileMode.NEW,
+        orig_name="",
+        new_name="LICENSE",
+        hunks=[],
+    )
+    out = renderers.build_full_file_diff(file)
+    assert "New empty file: LICENSE" in out or "No matching lines" in out
+
+
+def test_build_full_file_diff_none() -> None:
+    """
+    Should handle case when diff target is None.
+    """
+    out = renderers.build_full_file_diff(None)
+    assert "Diff target not found" in out or out == ""

xai_review-0.10.0/ai_review/tests/suites/services/review/inline/test_service.py

@@ -0,0 +1,101 @@
+from ai_review.services.review.inline.schema import InlineCommentListSchema
+from ai_review.services.review.inline.service import InlineCommentService
+
+
+def test_empty_output_returns_empty_list():
+    result = InlineCommentService.parse_model_output("")
+    assert isinstance(result, InlineCommentListSchema)
+    assert result.root == []
+
+
+def test_valid_json_array_parsed():
+    json_output = '[{"file": "a.py", "line": 1, "message": "use f-string"}]'
+    result = InlineCommentService.parse_model_output(json_output)
+    assert len(result.root) == 1
+    assert result.root[0].file == "a.py"
+    assert result.root[0].line == 1
+    assert result.root[0].message == "use f-string"
+
+
+def test_json_inside_code_block_parsed():
+    output = """```json
+    [
+      {"file": "b.py", "line": 42, "message": "check for None"}
+    ]
+    ```"""
+    result = InlineCommentService.parse_model_output(output)
+    assert len(result.root) == 1
+    assert result.root[0].file == "b.py"
+    assert result.root[0].line == 42
+
+
+def test_non_json_but_array_inside_text():
+    output = "some explanation...\n[ {\"file\": \"c.py\", \"line\": 7, \"message\": \"fix this\"} ]\nend"
+    result = InlineCommentService.parse_model_output(output)
+    assert len(result.root) == 1
+    assert result.root[0].file == "c.py"
+    assert result.root[0].line == 7
+
+
+def test_invalid_json_array_logs_and_returns_empty():
+    output = '[{"file": "d.py", "line": "oops", "message": "bad"}]'
+    result = InlineCommentService.parse_model_output(output)
+    assert result.root == []
+
+
+def test_no_json_array_found_logs_and_returns_empty():
+    output = "this is not json at all"
+    result = InlineCommentService.parse_model_output(output)
+    assert result.root == []
+
+
+def test_json_with_raw_newline_sanitized():
+    output = '[{"file": "e.py", "line": 3, "message": "line1\nline2"}]'
+    result = InlineCommentService.parse_model_output(output)
+    assert len(result.root) == 1
+    assert result.root[0].message == "line1\nline2"
+
+
+def test_json_with_tab_character_sanitized():
+    output = '[{"file": "f.py", "line": 4, "message": "a\tb"}]'
+    result = InlineCommentService.parse_model_output(output)
+    assert len(result.root) == 1
+    assert result.root[0].message == "a\tb"
+
+
+def test_json_with_null_byte_sanitized():
+    raw = "abc\0def"
+    output = f'[{{"file": "g.py", "line": 5, "message": "{raw}"}}]'
+    result = InlineCommentService.parse_model_output(output)
+    assert len(result.root) == 1
+    assert result.root[0].message == "abc\0def"
+
+
+def test_json_with_multiple_control_chars():
+    raw = "x\n\ry\t\0z"
+    output = f'[{{"file": "h.py", "line": 6, "message": "{raw}"}}]'
+    result = InlineCommentService.parse_model_output(output)
+    assert len(result.root) == 1
+    assert result.root[0].message == "x\n\ry\t\0z"
+
+
+def test_try_parse_valid_json():
+    raw = '[{"file": "ok.py", "line": 1, "message": "all good"}]'
+    result = InlineCommentService.try_parse_model_output(raw)
+    assert isinstance(result, InlineCommentListSchema)
+    assert len(result.root) == 1
+    assert result.root[0].file == "ok.py"
+
+
+def test_try_parse_needs_sanitization():
+    raw = '[{"file": "bad.py", "line": 2, "message": "line1\nline2"}]'
+    result = InlineCommentService.try_parse_model_output(raw)
+    assert result is not None
+    assert result.root[0].file == "bad.py"
+    assert "line1" in result.root[0].message
+
+
+def test_try_parse_totally_invalid_returns_none():
+    raw = "this is not json at all"
+    result = InlineCommentService.try_parse_model_output(raw)
+    assert result is None

{xai_review-0.7.0 → xai_review-0.10.0}/pyproject.toml

@@ -5,7 +5,7 @@ build-backend = "setuptools.build_meta"
 [project]
 name = "xai-review"
 readme = "README.md"
-version = "0.7.0"
+version = "0.10.0"
 license = { text = "Apache-2.0" }
 authors = [
     { name = "Nikita Filonov", email = "nikita.filonov@example.com" }

{xai_review-0.7.0 → xai_review-0.10.0}/xai_review.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: xai-review
-Version: 0.7.0
+Version: 0.10.0
 Summary: AI-powered code review tool
 Author-email: Nikita Filonov <nikita.filonov@example.com>
 Maintainer-email: Nikita Filonov <nikita.filonov@example.com>
@@ -298,6 +298,6 @@ ai-review:
 
 ## 📂 Examples
 
-- [./docs/ci/gitlab.yaml](./docs/ci/gitlab.yaml) — ready-to-use CI snippet
+- [./docs/ci](./docs/ci) — ready-to-use CI snippets
 - [./docs/configs](./docs/configs) — sample `.yaml`, `.json`, `.env` configs
 - [./docs/prompts](./docs/prompts) — prompt templates for Python/Go (light & strict modes)

{xai_review-0.7.0 → xai_review-0.10.0}/xai_review.egg-info/SOURCES.txt

@@ -29,6 +29,7 @@ ai_review/clients/openai/__init__.py
 ai_review/clients/openai/client.py
 ai_review/clients/openai/schema.py
 ai_review/libs/__init__.py
+ai_review/libs/json.py
 ai_review/libs/logger.py
 ai_review/libs/resources.py
 ai_review/libs/asynchronous/__init__.py
@@ -133,6 +134,7 @@ ai_review/tests/suites/clients/openai/__init__.py
 ai_review/tests/suites/clients/openai/test_client.py
 ai_review/tests/suites/clients/openai/test_schema.py
 ai_review/tests/suites/libs/__init__.py
+ai_review/tests/suites/libs/test_json.py
 ai_review/tests/suites/libs/config/__init__.py
 ai_review/tests/suites/libs/config/test_prompt.py
 ai_review/tests/suites/libs/diff/__init__.py

xai_review-0.7.0/ai_review/services/review/inline/service.py

@@ -1,38 +0,0 @@
-import re
-
-from pydantic import ValidationError
-
-from ai_review.libs.logger import get_logger
-from ai_review.services.review.inline.schema import InlineCommentListSchema
-
-logger = get_logger("INLINE_COMMENT_SERVICE")
-
-FIRST_JSON_ARRAY_RE = re.compile(r"\[[\s\S]*]", re.MULTILINE)
-CLEAN_JSON_BLOCK_RE = re.compile(r"```(?:json)?(.*?)```", re.DOTALL | re.IGNORECASE)
-
-
-class InlineCommentService:
-    @classmethod
-    def parse_model_output(cls, output: str) -> InlineCommentListSchema:
-        output = (output or "").strip()
-        if not output:
-            logger.warning("️LLM returned empty string for inline review")
-            return InlineCommentListSchema(root=[])
-
-        if match := CLEAN_JSON_BLOCK_RE.search(output):
-            output = match.group(1).strip()
-
-        try:
-            return InlineCommentListSchema.model_validate_json(output)
-        except ValidationError:
-            logger.warning("LLM output is not valid JSON, trying to extract first JSON array...")
-
-        if json_array_match := FIRST_JSON_ARRAY_RE.search(output):
-            try:
-                return InlineCommentListSchema.model_validate_json(json_array_match.group(0))
-            except ValidationError:
-                logger.exception("JSON array found but still invalid")
-        else:
-            logger.exception("No JSON array found in LLM output")
-
-        return InlineCommentListSchema(root=[])

xai_review-0.7.0/ai_review/tests/suites/services/review/inline/test_service.py

@@ -1,49 +0,0 @@
-from ai_review.services.review.inline.schema import InlineCommentListSchema
-from ai_review.services.review.inline.service import InlineCommentService
-
-
-def test_empty_output_returns_empty_list():
-    result = InlineCommentService.parse_model_output("")
-    assert isinstance(result, InlineCommentListSchema)
-    assert result.root == []
-
-
-def test_valid_json_array_parsed():
-    json_output = '[{"file": "a.py", "line": 1, "message": "use f-string"}]'
-    result = InlineCommentService.parse_model_output(json_output)
-    assert len(result.root) == 1
-    assert result.root[0].file == "a.py"
-    assert result.root[0].line == 1
-    assert result.root[0].message == "use f-string"
-
-
-def test_json_inside_code_block_parsed():
-    output = """```json
-    [
-      {"file": "b.py", "line": 42, "message": "check for None"}
-    ]
-    ```"""
-    result = InlineCommentService.parse_model_output(output)
-    assert len(result.root) == 1
-    assert result.root[0].file == "b.py"
-    assert result.root[0].line == 42
-
-
-def test_non_json_but_array_inside_text():
-    output = "some explanation...\n[ {\"file\": \"c.py\", \"line\": 7, \"message\": \"fix this\"} ]\nend"
-    result = InlineCommentService.parse_model_output(output)
-    assert len(result.root) == 1
-    assert result.root[0].file == "c.py"
-    assert result.root[0].line == 7
-
-
-def test_invalid_json_array_logs_and_returns_empty():
-    output = '[{"file": "d.py", "line": "oops", "message": "bad"}]'
-    result = InlineCommentService.parse_model_output(output)
-    assert result.root == []
-
-
-def test_no_json_array_found_logs_and_returns_empty():
-    output = "this is not json at all"
-    result = InlineCommentService.parse_model_output(output)
-    assert result.root == []