doocus first ver

doocus/extractors/tabular.py

@@ -0,0 +1,57 @@
+"""
+Tabular-family extractor: csv.
+
+Renders to a markdown table (capped preview) with row/column metadata. Uses only
+the stdlib so it always works.
+"""
+import csv
+from pathlib import Path
+
+from .base import ExtractionResult
+
+EXTRACTOR_ID = "tabular/csv@1"
+MAX_PREVIEW_ROWS = 200
+
+
+def _md_table(rows: list) -> str:
+    if not rows:
+        return ""
+    width = max(len(r) for r in rows)
+    rows = [r + [""] * (width - len(r)) for r in rows]
+    header, body = rows[0], rows[1:]
+
+    def esc(c: str) -> str:
+        return str(c).replace("|", "\\|").replace("\n", " ")
+
+    lines = ["| " + " | ".join(esc(c) for c in header) + " |",
+             "| " + " | ".join(["---"] * width) + " |"]
+    for r in body:
+        lines.append("| " + " | ".join(esc(c) for c in r) + " |")
+    return "\n".join(lines) + "\n"
+
+
+def extract(source: Path, options: dict) -> ExtractionResult:
+    warnings = []
+    with open(source, newline="", encoding="utf-8", errors="replace") as f:
+        sample = f.read(8192)
+        f.seek(0)
+        try:
+            dialect = csv.Sniffer().sniff(sample) if sample.strip() else csv.excel
+        except csv.Error:
+            dialect = csv.excel
+        rows = list(csv.reader(f, dialect))
+
+    total = len(rows)
+    cols = max((len(r) for r in rows), default=0)
+    preview = rows[: MAX_PREVIEW_ROWS + 1]  # +1 for header
+    if total > MAX_PREVIEW_ROWS + 1:
+        warnings.append(f"table has {total} rows; previewed first {MAX_PREVIEW_ROWS}")
+
+    content = _md_table(preview)
+    metadata = {
+        "rows": max(total - 1, 0),
+        "columns": cols,
+        "header": rows[0] if rows else [],
+    }
+    return ExtractionResult(content=content, metadata=metadata,
+                            warnings=warnings, extractor=EXTRACTOR_ID)