Merge pull request #8 from bcye/feature/modular-output-handlers

Basic setup of modular output handlers

.gitignore

@@ -11,3 +11,4 @@ wheels/
 
 .env
 node_modules
+output

output_handlers/__init__.py

@@ -0,0 +1,2 @@
+from .base_handler import BaseHandler
+from .filesystem import FilesystemHandler

output_handlers/base_handler.py

@@ -0,0 +1,54 @@
+"""Reference handler for output handlers."""
+from abc import ABC, abstractmethod
+import logging
+
+
+
+class BaseHandler(ABC):
+    """
+    Abstract base class for output handlers. Defines the standardized interface that all output handlers must implement.
+    In particular, it requires the implementation of an asynchronous ("private") method `_write_entry` to write a single entry to the output.
+    """
+
+    logger = logging.getLogger(__name__)
+
+    def __init__(self, fail_on_error: bool = True, **kwargs):
+        """
+        Initializes the BaseHandler with optional parameters.
+
+        Args:
+            fail_on_error (bool): If True, the handler will raise an exception on error. Defaults to True.
+            **kwargs: Additional keyword arguments for specific handler implementations.
+        """
+        self.fail_on_error = fail_on_error
+
+
+    @abstractmethod
+    async def _write_entry(self, entry: dict, uid: str) -> bool:
+        """
+        Asynchronously writes a single entry to the output. This method should gracefully handle any exceptions that may occur during the writing process and simply return False if an error occurs.
+
+        Args:
+            entry (dict): The entry to write (will be JSON-encoded).
+            uid (str): The unique identifier for the entry. The default id provided by wikivoyage is recommended. 
+        Returns:
+            bool: True if the entry was written successfully, False otherwise.
+        """
+        pass
+
+
+    async def write_entry(self, entry: dict, uid: str):
+        """
+        Public method to write an entry to the output. It handles exceptions and logs errors.
+
+        Args:
+            entry (dict): The entry to write (will be JSON-encoded).
+            uid (str): The unique identifier for the entry. The default id provided by wikivoyage is recommended. 
+        """
+        success = await self._write_entry(entry, uid)
+        if success:
+            self.logger.debug(f"Successfully wrote entry with UID {uid}")
+        else:
+            self.logger.error(f"Failed to write entry with UID {uid}")
+            if self.fail_on_error:
+                raise Exception(f"Failed to write entry with UID {uid}")

output_handlers/filesystem.py

@@ -0,0 +1,44 @@
+"""Handler that writes files to the filesystem."""
+from pathlib import Path
+import aiofiles
+from .base_handler import BaseHandler
+import json
+
+class FilesystemHandler(BaseHandler):
+    """
+    Handler that writes files to the filesystem.
+    """
+    def __init__(self, output_dir: str, **kwargs):
+        """
+        Initializes the FileSystemHandler with the specified output directory.
+
+        Args:
+            output_dir (str): The directory where files will be written.
+            **kwargs: Additional keyword arguments for the BaseHandler.
+        """
+        super().__init__(**kwargs)
+        self.output_dir = Path(output_dir)
+        # Ensure the target directory exists
+        self.output_dir.mkdir(parents=True, exist_ok=True)
+        self.logger.info(f"Output directory set to {self.output_dir}")
+
+
+    async def _write_entry(self, entry: dict, uid: str) -> bool:
+        """
+        Asynchronously writes a single entry to the filesystem.
+
+        Args:
+            entry (dict): The entry to write (will be JSON-encoded).
+            uid (str): The unique identifier for the entry.
+
+        Returns:
+            bool: True if the entry was written successfully, False otherwise.
+        """
+        try:
+            file_path = self.output_dir / f"{uid}.json"
+            async with aiofiles.open(file_path, 'w') as f:
+                await f.write(json.dumps(entry))
+            return True
+        except IOError as e:
+            self.logger.error(f"Error writing entry {uid}: {e}")
+            return False

pyproject.toml

@@ -5,6 +5,8 @@ description = "Add your description here"
 readme = "README.md"
 requires-python = ">=3.12"
 dependencies = [
+    "aiofiles>=24.1.0",
     "mwparserfromhell>=0.6.6",
+    "python-dotenv>=1.1.0",
     "wikitextparser>=0.56.3",
 ]

transform-documents.py

@@ -6,6 +6,11 @@ from typing import Dict, List, Any, Optional, Union, Tuple
 import os
 from pathlib import Path
 import sys
+import asyncio
+import importlib
+import logging
+from dotenv import load_dotenv
+load_dotenv()
 
 class WikivoyageParser:
     def __init__(self):
@@ -342,41 +347,120 @@ class WikivoyageParser:
             
         return json.dumps(root, indent=indent)
 
-def process_file(input_file: Path, parser: WikivoyageParser) -> None:
-    """Process a single wiki file and save JSON output"""
-    # Create output path with .json extension
-    output_file = input_file.with_suffix('.json')
+async def process_file(
+    input_file: Path,
+    parser: WikivoyageParser,
+    handler,
+) -> None:
+    """
+    Parse one wiki file and hand the resulting entry off to our handler.
+    Uses the filename (sans suffix) as the unique UID.
+    """
     
-    # Ensure output directory exists
-    output_file.parent.mkdir(parents=True, exist_ok=True)
+    text = input_file.read_text(encoding="utf-8")
+    entry = parser.parse(text)  # assume returns a dict
+    uid = input_file.stem
 
+    await handler.write_entry(entry, uid)
+
+def gather_handler_kwargs(handler_name: str) -> dict:
+    """
+    Find all ENV vars starting with HANDLER_<NAME>_ and turn them into kwargs.
+    E.g. HANDLER_SFTP_HOST=foo → {"host": "foo"}, HANDLER_SFTP_PORT=2222 → {"port": 2222}
+    """
+    prefix = f"HANDLER_{handler_name.upper()}_"
+    kwargs = {}
+
+    for env_key, val in os.environ.items():
+        if not env_key.startswith(prefix):
+            continue
+        param = env_key[len(prefix) :].lower()
+        # try to cast ints
+        if val.isdigit():
+            val = int(val)
+        kwargs[param] = val
+
+    return kwargs
+
+async def main():
+    logging.basicConfig(level=logging.DEBUG)
+
+    # 1. Which handler to load?
+    handler_name = os.getenv("HANDLER")
+    if not handler_name:
+        print("Error: set ENV HANDLER (e.g. 'filesystem')")
+        sys.exit(1)
+
+    # 2. Dynamic import
+    module_path = f"output_handlers.{handler_name}"
     try:
-        # Read and parse input file
-        with open(input_file, 'r', encoding='utf-8') as f:
-            wikitext = f.read()
+        mod = importlib.import_module(module_path)
+    except ImportError as e:
+        print(f"Error loading handler module {module_path}: {e}")
+        sys.exit(1)
 
-        result = parser.parse(wikitext)
+    # 3. Find the class: e.g. "sftp" → "SftpHandler"
+    class_name = handler_name.title().replace("_", "") + "Handler"
+    if not hasattr(mod, class_name):
+        print(f"{module_path} defines no class {class_name}")
+        sys.exit(1)
+    HandlerCls = getattr(mod, class_name)
 
-        # Write JSON output
-        with open(output_file, 'w', encoding='utf-8') as f:
-            f.write(parser.export_json())
+    # 4. Build kwargs from ENV
+    handler_kwargs = gather_handler_kwargs(handler_name)
 
-    except Exception as e:
-        print(f"Error processing {input_file}: {e}")
+    # 5. Instantiate
+    handler = HandlerCls(**handler_kwargs)
 
-def main():
-    # Initialize parser once for reuse
+    # 6. Prepare parser
     parser = WikivoyageParser()
 
-    # Get input directory from command line or use current directory
-    input_dir = Path(sys.argv[1] if len(sys.argv) > 1 else '.')
+    # 7. Which dir to walk?
+    input_dir = Path(sys.argv[1]) if len(sys.argv) > 1 else Path(".")
+    txt_files = list(input_dir.rglob("*.txt"))
 
-    # Process all .txt files recursively
-    for txt_file in input_dir.rglob('*.txt'):
-        print(f"Processing {txt_file}")
-        process_file(txt_file, parser)
+    if not txt_files:
+        print(f"No .txt files found under {input_dir}")
+    else:
+        for txt in txt_files:
+            await process_file(txt, parser, handler)
+
+    # 7) read concurrency setting
+    try:
+        max_conc = int(os.getenv("MAX_CONCURRENT", "0"))
+    except ValueError:
+        print("Error: MAX_CONCURRENT must be an integer")
+        sys.exit(1)
+
+    if max_conc < 0:
+        print("Error: MAX_CONCURRENT must be >= 0")
+        sys.exit(1)
+
+    # 8) schedule tasks
+    if max_conc == 0:
+        # unbounded
+        tasks = [
+            asyncio.create_task(process_file(txt, parser, handler))
+            for txt in txt_files
+        ]
+    else:
+        # bounded by semaphore
+        sem = asyncio.Semaphore(max_conc)
+
+        async def bounded(txt):
+            async with sem:
+                return await process_file(txt, parser, handler)
+
+        tasks = [
+            asyncio.create_task(bounded(txt))
+            for txt in txt_files
+        ]
+
+    # 9) run them all
+    await asyncio.gather(*tasks)
+
+    print("All done.")
 
-    print("Processing complete")
 
 if __name__ == "__main__":
-    main()
+    asyncio.run(main())

uv.lock

@@ -1,18 +1,31 @@
 version = 1
 requires-python = ">=3.12"
 
+[[package]]
+name = "aiofiles"
+version = "24.1.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/0b/03/a88171e277e8caa88a4c77808c20ebb04ba74cc4681bf1e9416c862de237/aiofiles-24.1.0.tar.gz", hash = "sha256:22a075c9e5a3810f0c2e48f3008c94d68c65d763b9b03857924c99e57355166c", size = 30247 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/a5/45/30bb92d442636f570cb5651bc661f52b610e2eec3f891a5dc3a4c3667db0/aiofiles-24.1.0-py3-none-any.whl", hash = "sha256:b4ec55f4195e3eb5d7abd1bf7e061763e864dd4954231fb8539a0ef8bb8260e5", size = 15896 },
+]
+
 [[package]]
 name = "mapvoyage-extract"
 version = "0.1.0"
 source = { virtual = "." }
 dependencies = [
+    { name = "aiofiles" },
     { name = "mwparserfromhell" },
+    { name = "python-dotenv" },
     { name = "wikitextparser" },
 ]
 
 [package.metadata]
 requires-dist = [
+    { name = "aiofiles", specifier = ">=24.1.0" },
     { name = "mwparserfromhell", specifier = ">=0.6.6" },
+    { name = "python-dotenv", specifier = ">=1.1.0" },
     { name = "wikitextparser", specifier = ">=0.56.3" },
 ]
 
@@ -29,6 +42,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/78/09/24c2f37524a3ebc3574975766748c7e4423ecefaa815c9fc4a324cbcf94a/mwparserfromhell-0.6.6-cp312-cp312-win_amd64.whl", hash = "sha256:cdc46c115b2495d4025920b7b30a6885a96d2b797ccc4009bf3cc02940ae55d3", size = 101071 },
 ]
 
+[[package]]
+name = "python-dotenv"
+version = "1.1.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/88/2c/7bb1416c5620485aa793f2de31d3df393d3686aa8a8506d11e10e13c5baf/python_dotenv-1.1.0.tar.gz", hash = "sha256:41f90bc6f5f177fb41f53e87666db362025010eb28f60a01c9143bfa33a2b2d5", size = 39920 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/1e/18/98a99ad95133c6a6e2005fe89faedf294a748bd5dc803008059409ac9b1e/python_dotenv-1.1.0-py3-none-any.whl", hash = "sha256:d7c01d9e2293916c18baf562d95698754b0dbbb5e74d457c45d4f6561fb9d55d", size = 20256 },
+]
+
 [[package]]
 name = "regex"
 version = "2024.11.6"