any-file to PDF (Phase 1): converters/ dispatcher, images/text/md/csv/json/xml/svg/pdf-passthrough, 112 tests green

converters/__init__.py

@@ -0,0 +1,85 @@
+"""converters — PHASE 1 multi-format dispatch (any common file -> PDF).
+
+Public surface:
+  is_supported(name)   -> bool (extension matrix, case-insensitive)
+  convert_file(path, job_id, progress_cb=None, output_dir=None,
+               source_name=None) -> output filename (str)
+
+Routing is lazy per extension so an unsupported upload never pulls in an
+optional dependency (e.g. PIL) it would not need. .epub routes to the
+existing EPUB pipeline (app.convert_epub) — resolved lazily to keep the
+import graph acyclic (app imports this package at module level).
+
+Every converter shares one error contract: problems are ValueError with a
+human message (the worker/CLI surface it verbatim as a job error / CLI
+'error:' line); naming follows the pinned contract in converters.common.
+"""
+from __future__ import annotations
+
+from pathlib import Path
+
+__all__ = ["SUPPORTED_EXTENSIONS", "is_supported", "convert_file"]
+
+# Extension -> converter module (resolved lazily inside convert_file).
+_ROUTING: dict[str, str] = {
+    ".html": "html_converter",
+    ".htm": "html_converter",
+    ".txt": "text_converter",
+    ".log": "text_converter",
+    ".md": "md_converter",
+    ".markdown": "md_converter",
+    ".csv": "csv_converter",
+    ".tsv": "csv_converter",
+    ".json": "json_converter",
+    ".xml": "xml_converter",
+    ".svg": "svg_converter",
+    ".png": "image_converter",
+    ".jpg": "image_converter",
+    ".jpeg": "image_converter",
+    ".webp": "image_converter",
+    ".tiff": "image_converter",
+    ".bmp": "image_converter",
+    ".pdf": "pdf_passthrough",
+}
+
+# .epub is supported (existing pipeline) but routes differently.
+SUPPORTED_EXTENSIONS = frozenset(_ROUTING) | {".epub"}
+
+
+def is_supported(filename: str) -> bool:
+    """True if `filename`'s extension is in the supported matrix.
+
+    Case-insensitive; extensionless names are unsupported.
+    """
+    if not filename:
+        return False
+    ext = Path(filename).suffix.lower()
+    return ext in SUPPORTED_EXTENSIONS
+
+
+def convert_file(path: Path, job_id: str, progress_cb=None, output_dir=None,
+                 source_name=None) -> str:
+    """Convert `path` to a PDF; return the output file name (no dir).
+
+    `source_name` is the original upload/CLI file name — the on-disk name
+    for web uploads is `{job_id}<ext>`, so the export name must be derived
+    from the original name, not the temp path.
+
+    Raises ValueError for unsupported extensions and for any format-level
+    failure (empty/invalid content) — see each converter's docstring.
+    """
+    ext = Path(path).suffix.lower()
+    if ext not in SUPPORTED_EXTENSIONS:
+        raise ValueError(
+            f"Unsupported file type: '{ext or '<none>'}'. "
+            f"Supported: {', '.join(sorted(SUPPORTED_EXTENSIONS))}"
+        )
+    if ext == ".epub":
+        import app  # lazy: the EPUB pipeline lives in app.py (avoid cycle)
+
+        return app.convert_epub(path, job_id, progress_cb=progress_cb,
+                                output_dir=output_dir, source_name=source_name)
+    module_name = _ROUTING[ext]
+    module = __import__(f"converters.{module_name}", fromlist=["convert"])
+    return module.convert(Path(path), job_id, progress_cb=progress_cb,
+                          output_dir=output_dir, source_name=source_name)