ocr-postprocess 0.1.0__tar.gz → 0.1.2__tar.gz

{ocr_postprocess-0.1.0 → ocr_postprocess-0.1.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ocr-postprocess
-Version: 0.1.0
+Version: 0.1.2
 Summary: Biến raw OCR text thành structured document — trích xuất trường dữ liệu, xử lý nhiễu, cross-check, render JSON/Markdown.
 Home-page: https://github.com/ohmygodvt95/ocr-postprocess
 Author: ohmygodvt95

{ocr_postprocess-0.1.0 → ocr_postprocess-0.1.2}/ocr_postprocess/cli.py

@@ -32,18 +32,22 @@ def classify(
     input_file: Optional[Path] = typer.Argument(
         None, help="Path to raw text file. Reads stdin if omitted."
     ),
-    profiles_dir: Path = typer.Option(
-        Path("profiles"), "--profiles", "-p", help="Profiles directory."
+    profiles_dir: Optional[Path] = typer.Option(
+        None, "--profiles", "-p", help="Profiles directory. Defaults to bundled profiles."
     ),
-    log_level: str = typer.Option("info", "--log-level", "-l"),
+    log_level: str = typer.Option("warning", "--log-level", "-l"),
 ) -> None:
     """Classify a document and print the best matching profile ID."""
     _setup_logging(log_level)
     from ocr_postprocess.pipeline import Pipeline
 
     raw = _read_input(input_file)
-    pipeline = Pipeline.from_default(profiles_dir=str(profiles_dir))
-    profile_id, score = pipeline.classify(raw)
+    try:
+        pipeline = Pipeline.from_default(profiles_dir=str(profiles_dir) if profiles_dir else None)
+        profile_id, score = pipeline.classify(raw)
+    except Exception as exc:
+        typer.echo(f"Error: {exc}", err=True)
+        raise typer.Exit(code=1)
     typer.echo(f"{profile_id}\t{score:.4f}")
 
 
@@ -52,14 +56,14 @@ def process(
     input_file: Optional[Path] = typer.Argument(
         None, help="Path to raw text file. Reads stdin if omitted."
     ),
-    profiles_dir: Path = typer.Option(
-        Path("profiles"), "--profiles", "-p", help="Profiles directory."
+    profiles_dir: Optional[Path] = typer.Option(
+        None, "--profiles", "-p", help="Profiles directory. Defaults to bundled profiles."
     ),
     output_format: str = typer.Option(
         "json", "--format", "-f", help="Output format: json, markdown, or llm."
     ),
     debug: bool = typer.Option(False, "--debug", "-d", help="Include debug trace."),
-    log_level: str = typer.Option("info", "--log-level", "-l"),
+    log_level: str = typer.Option("warning", "--log-level", "-l"),
 ) -> None:
     """Process a document and print extracted fields."""
     _setup_logging(log_level)
@@ -69,8 +73,12 @@ def process(
     from ocr_postprocess.renderer.markdown import render_markdown
 
     raw = _read_input(input_file)
-    pipeline = Pipeline.from_default(profiles_dir=str(profiles_dir))
-    doc = pipeline.process(raw, debug=debug)
+    try:
+        pipeline = Pipeline.from_default(profiles_dir=profiles_dir)
+        doc = pipeline.process(raw, debug=debug)
+    except Exception as exc:
+        typer.echo(f"Error: {exc}", err=True)
+        raise typer.Exit(code=1)
 
     if output_format == "markdown":
         typer.echo(render_markdown(doc))
@@ -83,7 +91,7 @@ def process(
 @app.command("validate-profile")
 def validate_profile(
     profile_file: Path = typer.Argument(..., help="Path to YAML profile file."),
-    log_level: str = typer.Option("info", "--log-level", "-l"),
+    log_level: str = typer.Option("warning", "--log-level", "-l"),
 ) -> None:
     """Validate a YAML profile file and print errors if any."""
     _setup_logging(log_level)
@@ -100,10 +108,10 @@ def validate_profile(
 @app.command("dump-canonical")
 def dump_canonical(
     input_file: Optional[Path] = typer.Argument(None, help="Path to raw text file."),
-    profiles_dir: Path = typer.Option(
-        Path("profiles"), "--profiles", "-p", help="Profiles directory."
+    profiles_dir: Optional[Path] = typer.Option(
+        None, "--profiles", "-p", help="Profiles directory. Defaults to bundled profiles."
     ),
-    log_level: str = typer.Option("info", "--log-level", "-l"),
+    log_level: str = typer.Option("warning", "--log-level", "-l"),
 ) -> None:
     """Process document and dump canonical JSON to stdout."""
     _setup_logging(log_level)
@@ -111,18 +119,33 @@ def dump_canonical(
     from ocr_postprocess.renderer.json_renderer import to_json
 
     raw = _read_input(input_file)
-    pipeline = Pipeline.from_default(profiles_dir=str(profiles_dir))
-    doc = pipeline.process(raw)
+    try:
+        pipeline = Pipeline.from_default(profiles_dir=profiles_dir)
+        doc = pipeline.process(raw)
+    except Exception as exc:
+        typer.echo(f"Error: {exc}", err=True)
+        raise typer.Exit(code=1)
     typer.echo(to_json(doc))
 
 
 def _read_input(path: Optional[Path]) -> str:
     """Read text from file path or stdin; exit with code 1 if no input available."""
     if path is not None:
-        return path.read_text(encoding="utf-8")
+        if not path.exists():
+            typer.echo(f"Error: file not found: {path}", err=True)
+            raise typer.Exit(code=1)
+        text = path.read_text(encoding="utf-8")
+        if not text.strip():
+            typer.echo(f"Error: file is empty: {path}", err=True)
+            raise typer.Exit(code=1)
+        return text
     if not sys.stdin.isatty():
-        return sys.stdin.read()
-    typer.echo("Error: no input provided", err=True)
+        text = sys.stdin.read()
+        if not text.strip():
+            typer.echo("Error: empty input from stdin", err=True)
+            raise typer.Exit(code=1)
+        return text
+    typer.echo("Error: no input provided (pass a file path or pipe text via stdin)", err=True)
     raise typer.Exit(code=1)
 
 

{ocr_postprocess-0.1.0 → ocr_postprocess-0.1.2}/ocr_postprocess/engine/reconciler.py

@@ -98,7 +98,7 @@ def reconcile_stage(ctx: PipelineContext) -> None:
                     field_label = fdef.aliases[0]
             readable_vals = ", ".join(f'"{v}"' for v in seen_vals)
             warning_msg = f"Conflict: {field_label} — nhiều giá trị khác nhau: {readable_vals}"
-            logger.warning("Conflict for field '%s': %s", key, detail)
+            logger.debug("Conflict for field '%s': %s", key, detail)
             ctx.warnings.append(warning_msg)
             merged.append(
                 best.model_copy(

{ocr_postprocess-0.1.0 → ocr_postprocess-0.1.2}/ocr_postprocess/pipeline.py

@@ -24,10 +24,10 @@ class Pipeline:
     def __init__(
         self,
         stages: list[Stage],
-        profiles_dir: str | Path = "profiles",
+        profiles_dir: str | Path | None = None,
     ) -> None:
         self._stages = stages
-        self._profiles_dir = Path(profiles_dir)
+        self._profiles_dir = Path(profiles_dir) if profiles_dir is not None else _BUNDLED_PROFILES_DIR
         self._profiles: dict = {}
 
     @classmethod

ocr_postprocess-0.1.2/ocr_postprocess/profiles/cccd_qr.yml

@@ -0,0 +1,89 @@
+id: cccd_qr
+version: 1
+display_name: "Căn cước công dân (quét mã QR)"
+language: ["vi"]
+
+classify:
+  any_of:
+    - contains_any: ["mã QR in trên CCCD", "Zalo không chịu trách nhiệm"]
+    - all_of:
+        - contains_any: ["Số CCCD", "Số CMND"]
+        - contains_any: ["Ngày cấp CCCD"]
+
+denoise:
+  drop_lines:
+    contains_any:
+      - "Thông tin này được lấy từ mã QR"
+      - "Zalo không chịu trách nhiệm"
+  collapse_repeats: true
+
+reconstruct:
+  fuzzy_threshold: 0.85
+  split_glued_labels: true
+  rejoin_wrapped_lines: true
+
+extract:
+  - name: tieu_de
+    constant: "Kết quả quét mã QR Code"
+
+  - name: so_cccd
+    aliases: ["Số CCCD"]
+    extractor: line_after_label
+    required: true
+    fuzzy_label: false
+    transform:
+      - {op: replace, from_: " ", to: ""}
+
+  - name: so_cmnd
+    aliases: ["Số CMND"]
+    extractor: line_after_label
+    fuzzy_label: false
+    transform:
+      - {op: replace, from_: " ", to: ""}
+
+  - name: ho_va_ten
+    aliases: ["Họ và tên"]
+    extractor: line_after_label
+    required: true
+    fuzzy_label: false
+
+  - name: ngay_sinh
+    aliases: ["Ngày sinh"]
+    extractor: regex_after_label
+    pattern: '(\d{1,2}[/\-.\s]\d{1,2}[/\-.\s]\d{4})'
+    fuzzy_label: false
+    next_lines: 1
+    required: true
+    transform:
+      - {op: to_date}
+
+  - name: gioi_tinh
+    aliases: ["Giới tính", "Nữ", "Nam"]
+    extractor: gender_vn
+    required: true
+
+  - name: noi_thuong_tru
+    aliases: ["Nơi thường trú"]
+    extractor: text_until_next_label
+    fuzzy_label: false
+    stop_labels: ["Ngày sinh", "Ngày cấp CCCD", "Giới tính"]
+    required: true
+
+  - name: ngay_cap
+    aliases: ["Ngày cấp CCCD", "Ngày cấp"]
+    extractor: regex_after_label
+    pattern: '(\d{1,2}[/\-.\s]\d{1,2}[/\-.\s]\d{4})'
+    fuzzy_label: false
+    next_lines: 1
+    required: true
+    transform:
+      - {op: to_date}
+
+output:
+  markdown:
+    title: "Kết quả quét mã QR Code"
+    sections:
+      - heading: "Thông tin cá nhân"
+        fields: [so_cccd, so_cmnd, ho_va_ten, gioi_tinh, ngay_sinh]
+      - heading: "Địa chỉ & cấp"
+        fields: [noi_thuong_tru, ngay_cap]

{ocr_postprocess-0.1.0 → ocr_postprocess-0.1.2}/ocr_postprocess.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ocr-postprocess
-Version: 0.1.0
+Version: 0.1.2
 Summary: Biến raw OCR text thành structured document — trích xuất trường dữ liệu, xử lý nhiễu, cross-check, render JSON/Markdown.
 Home-page: https://github.com/ohmygodvt95/ocr-postprocess
 Author: ohmygodvt95

{ocr_postprocess-0.1.0 → ocr_postprocess-0.1.2}/ocr_postprocess.egg-info/SOURCES.txt

@@ -52,6 +52,7 @@ ocr_postprocess/extractors/structured/mrz_cccd.py
 ocr_postprocess/profiles/__init__.py
 ocr_postprocess/profiles/_generic.yml
 ocr_postprocess/profiles/cccd_2024.yml
+ocr_postprocess/profiles/cccd_qr.yml
 ocr_postprocess/profiles/dang_kiem.yml
 ocr_postprocess/profiles/loader.py
 ocr_postprocess/profiles/matcher.py

{ocr_postprocess-0.1.0 → ocr_postprocess-0.1.2}/setup.py

@@ -5,7 +5,7 @@ with open("README.md", encoding="utf-8") as f:
 
 setup(
     name="ocr-postprocess",
-    version="0.1.0",
+    version="0.1.2",
     description="Biến raw OCR text thành structured document — trích xuất trường dữ liệu, xử lý nhiễu, cross-check, render JSON/Markdown.",
     long_description=long_description,
     long_description_content_type="text/markdown",