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

When llama-parse converts table-like sections to tables, the bBox property seems to use a default value which is not useful #300

New Issue
Closed
opened 2026-02-16 00:17:24 -05:00 by yindo · 3 comments
yindo commented 2026-02-16 00:17:24 -05:00
Owner
Copy Link
Copy Source

Originally created by @dan-cooke on GitHub (Oct 14, 2024).

Originally assigned to: @BinaryBrain on GitHub.

Describe the bug
This is a pretty niche one.

but when llama-parse identifies a "table-like" section in a document
image
In this example the section that starts:

Adviser fee: 0.50%
Platform fee: 0.30%
Underlying Fund costs: 0.22%

Is parsed as a table in the output:

  {
        "type": "table",
        "rows": [
          [
            "Fee Type",
            "Percentage"
          ],
          [
            "Adviser fee",
            "0.50%"
          ],
          [
            "Platform fee",
            "0.30%"
          ],
          [
            "Underlying Fund costs",
            "0.35%"
          ],
          [
            "Sarasin Discretionary Management Charge",
            "0.60%"
          ],
          [
            "Total Ongoing charges",
            "1.75%"
          ]
        ],

This is actually a great feature, i really like this.

The problem

When this parsing happens, it appears that the bBox property is no longer useful:

        "bBox": {
          "x": 0,
          "y": 0,
          "w": 595.3,
          "h": 841.89
        }

Workaround

I can pass a parsing instruction to say something like "Do not interpret table like structures as tables" but as I said... I actually really like this feature, so I'd rather not do that. Its much cleaner for our users to see a MD table instead of raw text for sections like this

Files
templi_J Bloggs & J Doe Illustration for New investments.docx

Job ID
1e0fa80c-b97f-4107-b720-ee6dd62845b8
You can find it here: https://cloud.llamaindex.ai/parse in the "History" tab.

Screenshots

Trying to draw a bounding box on my ui with this bug

image
As you can see one of the bounding boxes is (nearly correct) I believe the height is slightly wrong, but that is already raised #368, but the table parse is causing a box to be drawn around the entire document which is not quite right

Client:

  • Python Library
Originally created by @dan-cooke on GitHub (Oct 14, 2024). Originally assigned to: @BinaryBrain on GitHub. **Describe the bug** This is a pretty niche one. but when llama-parse identifies a "table-like" section in a document ![image](https://github.com/user-attachments/assets/be7b3269-cdb3-4e2c-836b-9eaf8b61443a) In this example the section that starts: ``` Adviser fee: 0.50% Platform fee: 0.30% Underlying Fund costs: 0.22% ``` Is parsed as a table in the output: ```.json { "type": "table", "rows": [ [ "Fee Type", "Percentage" ], [ "Adviser fee", "0.50%" ], [ "Platform fee", "0.30%" ], [ "Underlying Fund costs", "0.35%" ], [ "Sarasin Discretionary Management Charge", "0.60%" ], [ "Total Ongoing charges", "1.75%" ] ], ``` This is actually a great feature, i really like this. ## The problem When this parsing happens, it appears that the `bBox` property is no longer useful: ```.json "bBox": { "x": 0, "y": 0, "w": 595.3, "h": 841.89 } ``` ## Workaround I can pass a parsing instruction to say something like "Do not interpret table like structures as tables" but as I said... I actually really like this feature, so I'd rather not do that. Its much cleaner for our users to see a MD table instead of raw text for sections like this **Files** [templi_J Bloggs & J Doe Illustration for New investments.docx](https://github.com/user-attachments/files/17367120/templi_J.Bloggs.J.Doe.Illustration.for.New.investments.docx) **Job ID** 1e0fa80c-b97f-4107-b720-ee6dd62845b8 You can find it here: https://cloud.llamaindex.ai/parse in the "History" tab. **Screenshots** ### Trying to draw a bounding box on my ui with this bug ![image](https://github.com/user-attachments/assets/c42e461b-9e13-4227-a66e-43e5b9f629ae) As you can see one of the bounding boxes is (nearly correct) I believe the height is slightly wrong, but that is already raised #368, but the table parse is causing a box to be drawn around the entire document which is not quite right **Client:** - Python Library
yindo added the bug label 2026-02-16 00:17:24 -05:00
yindo closed this issue 2026-02-16 00:17:24 -05:00
yindo commented 2026-02-16 00:17:24 -05:00
Author
Owner
Copy Link
Copy Source

@hexapode commented on GitHub (Oct 14, 2024):

Great feedback thanks, this is actually a new feature will improve in coming weeks!

@hexapode commented on GitHub (Oct 14, 2024): Great feedback thanks, this is actually a new feature will improve in coming weeks!
yindo commented 2026-02-16 00:17:24 -05:00
Author
Owner
Copy Link
Copy Source

@zamedic commented on GitHub (Nov 21, 2024):

+1 to this issue.

Example
page_8

It starts ok, but as soon as it hits a table, things go, weird, suddenly, the Y is always 250, for everything.

page_8.jpg.json

@zamedic commented on GitHub (Nov 21, 2024): +1 to this issue. Example ![page_8](https://github.com/user-attachments/assets/f065db54-4cec-42b9-85b9-5291b3cc3ecd) It starts ok, but as soon as it hits a table, things go, weird, suddenly, the Y is always 250, for everything. [page_8.jpg.json](https://github.com/user-attachments/files/17847593/page_8.jpg.json)
yindo commented 2026-02-16 00:17:24 -05:00
Author
Owner
Copy Link
Copy Source

@BinaryBrain commented on GitHub (Dec 19, 2024):

This has finally been fixed!

@BinaryBrain commented on GitHub (Dec 19, 2024): This has finally been fixed!
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 bug
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#300
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?