run-llama/llama_cloud_services
1
0
Fork 0
mirror of https://github.com/run-llama/llama_cloud_services.git synced 2026-07-18 15:54:38 -04:00
Code Issues 352 Packages Projects Releases 10 Wiki Activity

Allow bounding_boxes and/or dpi settings for multimodal parsing with large pdf page #203

New Issue
Open
opened 2026-02-16 00:17:07 -05:00 by yindo · 2 comments
yindo commented 2026-02-16 00:17:07 -05:00
Owner
Copy Link
Copy Source

Originally created by @stonesthatwhisper on GitHub (Jul 24, 2024).

related: #233, #300.

Doing multimodal parsing with a large page, the texts may be scaled too small to be recognized well.
@hexapode does the image of the page respect the bounding_box setting? If it does I may be able to accomplish my goal by using multiple parser each with a different bounding_box. So,

feature request 1

Can we set a list of bounding_boxes to allow multiple samples of the same page to be sent conveniently to the model, and potentially return the results collectively?

feature request 2

Alternatively, could llama-parse allow a dpi settings for the sample image(s), and creating multiple samples of the big page within the bounding_box, and do the above under the hood?

Originally created by @stonesthatwhisper on GitHub (Jul 24, 2024). related: #233, #300. Doing multimodal parsing with a large page, the texts may be scaled too small to be recognized well. @hexapode does the image of the page respect the `bounding_box` setting? If it does I may be able to accomplish my goal by using multiple parser each with a different `bounding_box`. So, ### feature request 1 Can we set a list of `bounding_boxes` to allow multiple samples of the same page to be sent conveniently to the model, and potentially return the results collectively? ### feature request 2 Alternatively, could llama-parse allow a dpi settings for the sample image(s), and creating multiple samples of the big page within the `bounding_box`, and do the above under the hood?
yindo added the enhancement label 2026-02-16 00:17:07 -05:00
yindo commented 2026-02-16 00:17:08 -05:00
Author
Owner
Copy Link
Copy Source

@hexapode commented on GitHub (Jul 25, 2024):

The bounding_box parameter is also use for the mutlimodal pipeline. You can check at how the 'pages' are cropped by looking at the generated image (you can retrieve them by getting their id in the json output, then using the image endpoint)

@hexapode commented on GitHub (Jul 25, 2024): The `bounding_box` parameter is also use for the mutlimodal pipeline. You can check at how the 'pages' are cropped by looking at the generated image (you can retrieve them by getting their id in the json output, then using the image endpoint)
yindo commented 2026-02-16 00:17:08 -05:00
Author
Owner
Copy Link
Copy Source

@stonesthatwhisper commented on GitHub (Jul 25, 2024):

update: the .get_json_result() method shall give json. unfortunately, when I use this, I got

>> json_objs[0]['pages'][0]['images']
[{'name': 'page.jpg', 'height': 0, 'width': 0, 'x': 0, 'y': 0}]

@hexapode could you share an example where bounding_box is respected with gpt-4o?

Related: #300
Furthermore, the model starts to hallucinate about the texts in there. 95% of content it extracted is just wrong.

@stonesthatwhisper commented on GitHub (Jul 25, 2024): update: the `.get_json_result()` method shall give json. unfortunately, when I use this, I got ``` >> json_objs[0]['pages'][0]['images'] [{'name': 'page.jpg', 'height': 0, 'width': 0, 'x': 0, 'y': 0}] ``` @hexapode could you share an example where bounding_box is respected with gpt-4o? Related: #300 Furthermore, the model starts to hallucinate about the texts in there. 95% of content it extracted is just wrong.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
LlamaCloud
LlamaExtract
LlamaParse
bug
client
complex layout
documentation
duplicate
enhancement
figure
font
header footer
help wanted
ocr
parser
pdf_debug
pdf_debug
pull-request
Mirrored from GitHub Pull Request
 python-lib
question
server_issue
server_issue
ui
No Label enhancement
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: run-llama/llama_cloud_services#203
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?