From f2e9c0784c842612641171754ce51362e298088d Mon Sep 17 00:00:00 2001
From: Krishnan <krishkoushik@users.noreply.github.com>
Date: Tue, 20 May 2025 13:29:38 +0530
Subject: [PATCH] fix: incorrect force_backend_text behaviour for VLM DocTag
 pipelines (#1371)

* Fix force_backend_text

Signed-off-by: Krishnan Raghavan <krishnanraghavan@Krishnans-MacBook-Air.local>

* empty commit to retrigger CI

Signed-off-by: Panos Vagenas <pva@zurich.ibm.com>

---------

Signed-off-by: Krishnan Raghavan <krishnanraghavan@Krishnans-MacBook-Air.local>
Signed-off-by: Panos Vagenas <pva@zurich.ibm.com>
Co-authored-by: Krishnan Raghavan <krishnanraghavan@Krishnans-MacBook-Air.local>
Co-authored-by: Panos Vagenas <pva@zurich.ibm.com>
---
 docling/pipeline/vlm_pipeline.py | 36 +++++++++++++++-----------------
 1 file changed, 17 insertions(+), 19 deletions(-)

diff --git a/docling/pipeline/vlm_pipeline.py b/docling/pipeline/vlm_pipeline.py
index 9a7b51e..6605c1c 100644
--- a/docling/pipeline/vlm_pipeline.py
+++ b/docling/pipeline/vlm_pipeline.py
@@ -136,25 +136,23 @@ class VlmPipeline(PaginatedPipeline):
                 conv_res.document.load_from_doctags(doctags_doc)
 
                 # If forced backend text, replace model predicted text with backend one
-                if page.size:
-                    if self.force_backend_text:
-                        scale = self.pipeline_options.images_scale
-                        for element, _level in conv_res.document.iterate_items():
-                            if (
-                                not isinstance(element, TextItem)
-                                or len(element.prov) == 0
-                            ):
-                                continue
-                            crop_bbox = (
-                                element.prov[0]
-                                .bbox.scaled(scale=scale)
-                                .to_top_left_origin(
-                                    page_height=page.size.height * scale
-                                )
-                            )
-                            txt = self.extract_text_from_backend(page, crop_bbox)
-                            element.text = txt
-                            element.orig = txt
+                if self.force_backend_text:
+                    scale = self.pipeline_options.images_scale
+                    for element, _level in conv_res.document.iterate_items():
+                        if not isinstance(element, TextItem) or len(element.prov) == 0:
+                            continue
+                        page_ix = element.prov[0].page_no - 1
+                        page = conv_res.pages[page_ix]
+                        if not page.size:
+                            continue
+                        crop_bbox = (
+                            element.prov[0]
+                            .bbox.scaled(scale=scale)
+                            .to_top_left_origin(page_height=page.size.height * scale)
+                        )
+                        txt = self.extract_text_from_backend(page, crop_bbox)
+                        element.text = txt
+                        element.orig = txt
             elif (
                 self.pipeline_options.vlm_options.response_format
                 == ResponseFormat.MARKDOWN