From 465be4ca041b6f4eec5adc09c095417197ecabf0 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E2=80=9CNaeel=E2=80=9D?= <tazet@narod.ru>
Date: Sat, 27 Jun 2026 13:04:47 +0400
Subject: [PATCH] =?UTF-8?q?fix:=20PyPDF2=20=E2=86=92=20pdfplumber=20?=
 =?UTF-8?q?=E2=80=94=20=D1=81=D0=BE=D1=85=D1=80=D0=B0=D0=BD=D1=8F=D0=B5?=
 =?UTF-8?q?=D1=82=20=D1=81=D1=82=D1=80=D1=83=D0=BA=D1=82=D1=83=D1=80=D1=83?=
 =?UTF-8?q?=20=D1=82=D0=B0=D0=B1=D0=BB=D0=B8=D1=86=20=D0=B2=20PDF?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 deploy/services/parse.py | 34 ++++++++++++++++++++++++----------
 1 file changed, 24 insertions(+), 10 deletions(-)

diff --git a/deploy/services/parse.py b/deploy/services/parse.py
index 0576809..20cc37e 100644
--- a/deploy/services/parse.py
+++ b/deploy/services/parse.py
@@ -1,6 +1,5 @@
-"""Parse service — PDF (PyPDF2), DOCX (python-docx), plain text fallback."""
+"""Parse service — PDF (pdfplumber), DOCX (python-docx), plain text fallback."""
 import io
-from PyPDF2 import PdfReader
 
 
 def parse_file(filename, data):
@@ -20,15 +19,30 @@ def parse_file(filename, data):
 
 
 def _parse_pdf(data):
-    reader = PdfReader(io.BytesIO(data))
+    """Parse PDF with pdfplumber — preserves table structure (unlike PyPDF2)."""
+    import pdfplumber
     elements = []
-    for page in reader.pages:
-        text = page.extract_text()
-        if text:
-            for line in text.split("\n"):
-                line = line.strip()
-                if line:
-                    elements.append({"type": "paragraph", "text": line, "style": ""})
+    with pdfplumber.open(io.BytesIO(data)) as pdf:
+        for page in pdf.pages:
+            # Tables first — preserves column structure critical for specs
+            tables = page.extract_tables()
+            for table in tables:
+                if table:
+                    rows = []
+                    for row in table:
+                        if row:
+                            cells = [str(cell or "").strip() for cell in row]
+                            if any(cells):
+                                rows.append(cells)
+                    if rows:
+                        elements.append({"type": "table", "rows": rows})
+            # Remaining text as paragraphs
+            text = page.extract_text()
+            if text:
+                for line in text.split("\n"):
+                    line = line.strip()
+                    if line:
+                        elements.append({"type": "paragraph", "text": line, "style": ""})
     return {"status": "parsed", "element_count": len(elements), "elements": elements}