From 2290c2a2f0fc691900674451e2d4410da85c8956 Mon Sep 17 00:00:00 2001
From: H <43509927+guoyuhao2330@users.noreply.github.com>
Date: Thu, 11 Jul 2024 14:37:55 +0800
Subject: [PATCH] fix pdf_paser char content confusion (#1462)

### What problem does this PR solve?

#1407

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)
---
 deepdoc/parser/pdf_parser.py | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/deepdoc/parser/pdf_parser.py b/deepdoc/parser/pdf_parser.py
index fce9e91de..c7aaa9d58 100644
--- a/deepdoc/parser/pdf_parser.py
+++ b/deepdoc/parser/pdf_parser.py
@@ -287,14 +287,15 @@ class RAGFlowPdfParser:
         )
 
         # solve char content confusion
-        record_error_length = 0
+        record_error_length, ct = 0, 1
         for c in chars[0:128]:
             ii = Recognizer.find_overlapped(c, bxs)
             if ii is None:
                 continue
-            record_error_length += abs((bxs[ii]["bottom"] + bxs[ii]["bottom"] - c["bottom"] - c["top"]) / 2)
+            record_error_length += abs((bxs[ii]["bottom"] + bxs[ii]["top"] - c["bottom"] - c["top"]) / 2)
+            ct += 1
 
-        record_error_length = record_error_length / 128
+        record_error_length = record_error_length / ct
         for char in chars:
             char["top"] -= record_error_length
             char["bottom"] -= record_error_length