xai-review 0.8.0__py3-none-any.whl → 0.10.0__py3-none-any.whl

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)

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

ai_review/services/review/inline/service.py

@@ -2,6 +2,7 @@ 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
 
@@ -12,6 +13,19 @@ CLEAN_JSON_BLOCK_RE = re.compile(r"```(?:json)?(.*?)```", re.DOTALL | re.IGNOREC
 
 
 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()
@@ -22,17 +36,18 @@ class InlineCommentService:
         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 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):
-            try:
-                return InlineCommentListSchema.model_validate_json(json_array_match.group(0))
-            except ValidationError:
-                logger.exception("JSON array found but still invalid")
+            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.exception("No JSON array found in LLM output")
+            logger.error("No JSON array found in LLM output")
 
         return InlineCommentListSchema(root=[])

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"

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 == ""

ai_review/tests/suites/services/review/inline/test_service.py

@@ -47,3 +47,55 @@ 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-0.8.0.dist-info → xai_review-0.10.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: xai-review
-Version: 0.8.0
+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>

{xai_review-0.8.0.dist-info → xai_review-0.10.0.dist-info}/RECORD

@@ -26,6 +26,7 @@ ai_review/clients/openai/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZ
 ai_review/clients/openai/client.py,sha256=N07ZRnbptOtab8imMUZbGL-kRoOwIZmNYbHySumD3IU,1707
 ai_review/clients/openai/schema.py,sha256=glxwMtBrDA6W0BQgH-ruKe0bKH3Ps1P-Y1-2jGdqaUM,764
 ai_review/libs/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+ai_review/libs/json.py,sha256=koGnjcPgHBq3DHvzr090pM1ZCPtM9TjAoqWhnZJIo1I,460
 ai_review/libs/logger.py,sha256=LbXR2Zk1btJ-83I-vHee7cUETgT1mHToSsqEI_8uM0U,370
 ai_review/libs/resources.py,sha256=s9taAbL1Shl_GiGkPpkkivUcM1Yh6d_IQAG97gffsJU,748
 ai_review/libs/asynchronous/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -78,7 +79,7 @@ ai_review/services/cost/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG
 ai_review/services/cost/schema.py,sha256=K3uCIMMxGL8AaIPh4a-d0mT5uIJuk3f805DkP8o8DtY,1323
 ai_review/services/cost/service.py,sha256=rK-jw0lDszv_O13CRZAGK7R-fB-Y7xakX8aVb86zcEk,2103
 ai_review/services/diff/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-ai_review/services/diff/renderers.py,sha256=L9SEnkrYjt0kfoKl8qBH1dUDUrnA32CzCx0YI9IPvco,5611
+ai_review/services/diff/renderers.py,sha256=tEUml-uqsi5FNoU2NYjxehsZHU61dTPR2VnFi7QVzV4,5861
 ai_review/services/diff/schema.py,sha256=17GAQY1-ySwREJ1-NKNKgBcstMJ5Hb42FcFG2p7i6Rs,94
 ai_review/services/diff/service.py,sha256=FDuMw_y2QVQcfSbkVv3H2uGf1sIMeQ0_KHYCPhCU24g,3498
 ai_review/services/diff/tools.py,sha256=YHmH6Ult_rucCd563UhG0geMzqrPhqKFZKyug79xNuA,1963
@@ -102,7 +103,7 @@ ai_review/services/review/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJW
 ai_review/services/review/service.py,sha256=8YhRFqhZAk2pAnkDaytKSCENlOeOti1brAJq3R9tVMY,8394
 ai_review/services/review/inline/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ai_review/services/review/inline/schema.py,sha256=ry8sJdTgusQvFW51neRiapzgzVGwswwJzdYhaV3hbT0,1545
-ai_review/services/review/inline/service.py,sha256=2joeGCoPLacQAsmEtEKzmqPTQNYFYW8Dq3Cyfn3HZ_I,1408
+ai_review/services/review/inline/service.py,sha256=qJqtjLY1En07_ZiI_wnhJHZDFfUSUHDqkMXutd1ZuMc,2131
 ai_review/services/review/policy/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ai_review/services/review/policy/service.py,sha256=yGDePLxAEF3N1Pkh47jGVd-4dGEESyxDXIXxV7KQfuY,2027
 ai_review/services/review/summary/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -130,6 +131,7 @@ ai_review/tests/suites/clients/openai/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCe
 ai_review/tests/suites/clients/openai/test_client.py,sha256=Ox5ifP1C_gSeDRacBa9DnXhe__Z8-WcbzR2JH_V3xKo,1050
 ai_review/tests/suites/clients/openai/test_schema.py,sha256=x1tamS4GC9pOTpjieKDbK2D73CVV4BkATppytwMevLo,1599
 ai_review/tests/suites/libs/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+ai_review/tests/suites/libs/test_json.py,sha256=vmbkzRRyPuP-0uyLLlj-Tf_0MAWI_V2wDev9sM8tAHw,1054
 ai_review/tests/suites/libs/config/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ai_review/tests/suites/libs/config/test_prompt.py,sha256=kDMTnykC54tTPfE6cqYRBbV8d5jzAKucXdJfwtqUybM,2242
 ai_review/tests/suites/libs/diff/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -140,7 +142,7 @@ ai_review/tests/suites/libs/template/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeu
 ai_review/tests/suites/libs/template/test_render.py,sha256=n-ss5bd_hwc-RzYmqWmFM6KSlP1zLSnlsW1Yki12Bpw,1890
 ai_review/tests/suites/services/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ai_review/tests/suites/services/diff/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-ai_review/tests/suites/services/diff/test_renderers.py,sha256=XFXaOTGAdPOJRb8w-1BOX-eci1C49mKvH4vckybrJSg,5641
+ai_review/tests/suites/services/diff/test_renderers.py,sha256=IKOpsGedONNW8ZfYTAk0Vq0hfFi7L6TpWs8vVVQroj0,6273
 ai_review/tests/suites/services/diff/test_service.py,sha256=iFkGX9Vj2X44JU3eFsoHsg9o9353eKX-QCv_J9KxfzU,3162
 ai_review/tests/suites/services/diff/test_tools.py,sha256=HBQ3eCn-kLeb_k5iTgyr09x0VwLYSegSbxm0Qk9ZrCc,3543
 ai_review/tests/suites/services/prompt/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -149,15 +151,15 @@ ai_review/tests/suites/services/prompt/test_service.py,sha256=M8vvBhEbyHnXCSiIRu
 ai_review/tests/suites/services/review/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ai_review/tests/suites/services/review/inline/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ai_review/tests/suites/services/review/inline/test_schema.py,sha256=tIz-1UA_GgwcdsyUqgrodiiVVmd_jhoOVmtEwzRVWiY,2474
-ai_review/tests/suites/services/review/inline/test_service.py,sha256=5YP5xoijfn39jdAlmL5fEWBNjD4XhsgKnCPBUZDMZfw,1729
+ai_review/tests/suites/services/review/inline/test_service.py,sha256=ZcrEWKyD-Fsu3tqAOzT1Et1bxLbX6Hn0_Jc9xE_1_78,3566
 ai_review/tests/suites/services/review/policy/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ai_review/tests/suites/services/review/policy/test_service.py,sha256=kRWT550OjWYQ7ZfsihBRc-tx-NMkhlynEsqur55RK0M,3687
 ai_review/tests/suites/services/review/summary/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ai_review/tests/suites/services/review/summary/test_schema.py,sha256=xSoydvABZldHaVDa0OBFvYrj8wMuZqUDN3MO-XdvxOI,819
 ai_review/tests/suites/services/review/summary/test_service.py,sha256=8UMvi_NL9frm280vD6Q1NCDrdI7K8YbXzoViIus-I2g,541
-xai_review-0.8.0.dist-info/licenses/LICENSE,sha256=p-v8m7Kmz4KKc7PcvsGiGEmCw9AiSXY4_ylOPy_u--Y,11343
-xai_review-0.8.0.dist-info/METADATA,sha256=d7MEYbJjP_FPZpuZdbHmtaak6gyDbk72-BV07n3u67o,9617
-xai_review-0.8.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-xai_review-0.8.0.dist-info/entry_points.txt,sha256=JyC5URanMi5io5P_PXQf7H_I1OGIpk5cZQhaPQ0g4Zs,53
-xai_review-0.8.0.dist-info/top_level.txt,sha256=sTsZbfzLoqvRZKdKa-BcxWvjlHdrpbeJ6DrGY0EuR0E,10
-xai_review-0.8.0.dist-info/RECORD,,
+xai_review-0.10.0.dist-info/licenses/LICENSE,sha256=p-v8m7Kmz4KKc7PcvsGiGEmCw9AiSXY4_ylOPy_u--Y,11343
+xai_review-0.10.0.dist-info/METADATA,sha256=eiLnNDKQiqcKlXpaAL2-rkF8mHENCWJtUAUliMzeVic,9618
+xai_review-0.10.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+xai_review-0.10.0.dist-info/entry_points.txt,sha256=JyC5URanMi5io5P_PXQf7H_I1OGIpk5cZQhaPQ0g4Zs,53
+xai_review-0.10.0.dist-info/top_level.txt,sha256=sTsZbfzLoqvRZKdKa-BcxWvjlHdrpbeJ6DrGY0EuR0E,10
+xai_review-0.10.0.dist-info/RECORD,,