refactor: excel parse

server/parsers/cad_parser.py

@@ -0,0 +1,121 @@
+"""CAD parser — DXF via ezdxf, DWG via ODA File Converter."""
+from __future__ import annotations
+
+import os
+import subprocess
+import tempfile
+import ezdxf
+
+
+def parse_cad(file_path: str) -> dict:
+    ext = os.path.splitext(file_path)[1].lower()
+
+    if ext == ".dwg":
+        dxf_path = _convert_dwg(file_path)
+        if dxf_path is None:
+            return {"markdown": "", "error": "DWG 需要 ODA File Converter，下载: https://www.opendesign.com/guestfiles/oda_file_converter"}
+        file_path = dxf_path
+
+    try:
+        doc = ezdxf.readfile(file_path)
+    except Exception as e:
+        return {"markdown": "", "error": f"无法解析 DXF: {e}"}
+
+    return _extract(doc)
+
+
+def _convert_dwg(dwg_path: str) -> str | None:
+    candidates = [
+        "ODAFileConverter",
+        "/usr/local/bin/ODAFileConverter",
+        "/Applications/ODAFileConverter.app/Contents/MacOS/ODAFileConverter",
+        r"C:\Program Files\ODA\ODAFileConverter\ODAFileConverter.exe",
+    ]
+    converter = None
+    for c in candidates:
+        if os.path.isfile(c) or _which(c):
+            converter = c
+            break
+    if not converter:
+        return None
+
+    input_dir = os.path.dirname(os.path.abspath(dwg_path))
+    output_dir = tempfile.mkdtemp(prefix="engimind_cad_")
+    filename = os.path.basename(dwg_path)
+    try:
+        subprocess.run([converter, input_dir, output_dir, "ACAD2018", "DXF", "0", "1", filename],
+                       check=True, timeout=60, capture_output=True)
+    except Exception:
+        return None
+
+    base = os.path.splitext(filename)[0]
+    dxf = os.path.join(output_dir, base + ".dxf")
+    return dxf if os.path.isfile(dxf) else None
+
+
+def _which(name: str) -> bool:
+    try:
+        return subprocess.run(["which", name], capture_output=True, timeout=5).returncode == 0
+    except Exception:
+        return False
+
+
+def _extract(doc: ezdxf.document.Drawing) -> dict:
+    parts = ["## CAD 图纸解析结果\n"]
+
+    # Layers
+    layers = [{"name": l.dxf.name, "color": l.dxf.color} for l in doc.layers]
+    if layers:
+        parts.append("### 图层列表\n\n| 图层名 | 颜色编号 |\n| --- | --- |")
+        for l in layers:
+            parts.append(f"| {l['name']} | {l['color']} |")
+        parts.append("")
+
+    msp = doc.modelspace()
+    entity_count = {}
+    texts, dimensions, blocks = [], [], set()
+
+    for e in msp:
+        et = e.dxftype()
+        entity_count[et] = entity_count.get(et, 0) + 1
+        if et == "TEXT":
+            texts.append(e.dxf.text)
+        elif et == "MTEXT":
+            texts.append(e.text)
+        elif et == "DIMENSION":
+            try:
+                dimensions.append(e.dxf.text or "测量值")
+            except Exception:
+                pass
+        elif et == "INSERT":
+            blocks.add(e.dxf.name)
+
+    if entity_count:
+        parts.append("### 实体统计\n\n| 实体类型 | 数量 |\n| --- | --- |")
+        for et, cnt in sorted(entity_count.items()):
+            parts.append(f"| {et} | {cnt} |")
+        parts.append("")
+
+    if texts:
+        parts.append("### 文字标注\n")
+        for t in texts[:200]:
+            clean = t.strip().replace("\n", " ")
+            if clean:
+                parts.append(f"- {clean}")
+        if len(texts) > 200:
+            parts.append(f"\n> 共 {len(texts)} 条，仅显示前 200 条。")
+        parts.append("")
+
+    if dimensions:
+        parts.append("### 尺寸标注\n")
+        for d in dimensions[:100]:
+            parts.append(f"- {d}")
+        parts.append("")
+
+    if blocks:
+        parts.append("### 使用的图块\n")
+        for b in sorted(blocks):
+            parts.append(f"- {b}")
+        parts.append("")
+
+    return {"markdown": "\n".join(parts)}