run-llama/llama_cloud_services
1
0
Fork 0
mirror of https://github.com/run-llama/llama_cloud_services.git synced 2026-08-26 16:56:36 -04:00
Code Issues 352 Packages Projects Releases 10 Wiki Activity

NO_CONTENT_HERE - LLamaParse has STOPPED working with SCANNED PDFs #422

New Issue
Closed
opened 2026-02-16 00:17:48 -05:00 by yindo · 1 comment
yindo commented 2026-02-16 00:17:48 -05:00
Owner
Copy Link
Copy Source

Originally created by @Edwardmaster7 on GitHub (Feb 7, 2025).

Describe the bug
I'm working with scanned pdfs of an old periodic. Since yesterday Feb 2, I'm getting an error rate of nearly 99% on all pages processed, coming with the tag NO_CONTENT_HERE. Even on the playground I'm getting this recurrently.

Files
Image

Image

Job ID
498593a2-5020-4f21-a51c-4311e078e7d7

a64b55b1-380a-4fab-a611-8fcc8645f45b

Client:

  • Python Library
  • Frontend (cloud.llamaindex.ai)

Additional context
I'm using the current parameters, on both the api and playground:

api_key = os.getenv("LLAMA_CLOUD_API_KEY"), # LlamaParser API key
result_type = "markdown", # Result format: "json", "markdown" or "text"
verbose = verbose, # Enables detailed logs if needed
use_vendor_multimodal_model=True, # Enables multimodal model usage
vendor_multimodal_model_name = "gemini-1.5-pro", # Multimodal model to be used
vendor_multimodal_api_key = os.getenv("GEMINI_API_KEY"), # Multimodal model API key
page_prefix = "\n", # Prefix to be added before each page
page_suffix = "\n", # Suffix to be added after each page
page_separator = page_separator, # Custom page separator
complemental_formatting_instruction = complemental_formatting_instruction, # Complementary formatting instruction
content_guideline_instruction = content_guideline_instruction, # Content guideline instruction
do_not_cache=True,

complemental_formatting_instruction = """

  • Read and transcribe the file carefully from beginning to end, without skipping sections or lines.
  • Do not alter the original spelling of words.
  • Eliminate any line breaks within words.
  • Maintain the natural reading order as the text appears in the file.
  • Pay attention to maintaining grammatical and semantic coherence of the text.
  • Avoid random characters and noise as much as possible.
  • If there are any doubts or illegible words, use "[Illegible Text]".
  • Do not include page headers, focus only on the article texts."""
    content_guideline_instruction = """I am sending a file from an old periodical that contains text, possibly in multiple columns or pages, with potential presence of images and syllable breaks at the end of lines. Your task is to extract all the text, maintaining the spelling and natural reading order of the text."""
    prompt = """ROLE: You are a transcriber specialized in processing PDF files containing text, possibly in multiple columns or pages, with the presence of images and syllable breaks at the end of lines."""
Originally created by @Edwardmaster7 on GitHub (Feb 7, 2025). **Describe the bug** I'm working with scanned pdfs of an old periodic. Since yesterday Feb 2, I'm getting an error rate of nearly 99% on all pages processed, coming with the tag NO_CONTENT_HERE. Even on the playground I'm getting this recurrently. **Files** ![Image](https://github.com/user-attachments/assets/828b00d4-f37e-42ff-ba88-83978304dd39) ![Image](https://github.com/user-attachments/assets/88255e7f-fac5-45c4-b74f-1dab7f57f02b) **Job ID** 498593a2-5020-4f21-a51c-4311e078e7d7 a64b55b1-380a-4fab-a611-8fcc8645f45b **Client:** - Python Library - Frontend (cloud.llamaindex.ai) **Additional context** I'm using the current parameters, on both the api and playground: api_key = os.getenv("LLAMA_CLOUD_API_KEY"), # LlamaParser API key result_type = "markdown", # Result format: "json", "markdown" or "text" verbose = verbose, # Enables detailed logs if needed use_vendor_multimodal_model=True, # Enables multimodal model usage vendor_multimodal_model_name = "gemini-1.5-pro", # Multimodal model to be used vendor_multimodal_api_key = os.getenv("GEMINI_API_KEY"), # Multimodal model API key page_prefix = "\n", # Prefix to be added before each page page_suffix = "\n", # Suffix to be added after each page page_separator = page_separator, # Custom page separator complemental_formatting_instruction = complemental_formatting_instruction, # Complementary formatting instruction content_guideline_instruction = content_guideline_instruction, # Content guideline instruction do_not_cache=True, complemental_formatting_instruction = """ - Read and transcribe the file carefully from beginning to end, without skipping sections or lines. - Do not alter the original spelling of words. - Eliminate any line breaks within words. - Maintain the natural reading order as the text appears in the file. - Pay attention to maintaining grammatical and semantic coherence of the text. - Avoid random characters and noise as much as possible. - If there are any doubts or illegible words, use "[Illegible Text]". - Do not include page headers, focus only on the article texts.""" content_guideline_instruction = """I am sending a file from an old periodical that contains text, possibly in multiple columns or pages, with potential presence of images and syllable breaks at the end of lines. Your task is to extract all the text, maintaining the spelling and natural reading order of the text.""" prompt = """ROLE: You are a transcriber specialized in processing PDF files containing text, possibly in multiple columns or pages, with the presence of images and syllable breaks at the end of lines."""
yindo added the bug label 2026-02-16 00:17:48 -05:00
yindo closed this issue 2026-02-16 00:17:48 -05:00
yindo commented 2026-02-16 00:17:48 -05:00
Author
Owner
Copy Link
Copy Source

@BinaryBrain commented on GitHub (Feb 7, 2025):

Hi @Edwardmaster7,
Gemini 1.5 Flash and Pro are sending a lot of 503 errors and there's little we can do about it.
I recommend you try to set vendor_multimodal_model_name="gemini-2.0-flash-001" as the quality is overall way better and the model is also cheaper.

@BinaryBrain commented on GitHub (Feb 7, 2025): Hi @Edwardmaster7, Gemini 1.5 Flash and Pro are sending a lot of 503 errors and there's little we can do about it. I recommend you try to set `vendor_multimodal_model_name="gemini-2.0-flash-001"` as the quality is overall way better and the model is also cheaper.
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#422
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?