Page numbers returned in response

[stevelizcano]

For some indexing into our VectorDB, it's very helpful to know the page numbers if possible.

Is this feature possible now, or to add?

Either way, love the library. Great work!

[anoopshrma]

Currently page number is not returned in the metadata. But page content is separated using "-----------". You can separate the text based on the string pattern and provide page number for that by yourself.

PS: Adding more info in metadata is being implemented and will be available soon.

[hexapode]

We will provide a more reliable way to get page number (JSON output) in a future release

[Laktus]

@hexapode When would such a release come? Until then is @anoopshrma suggestion a valid solution? Can we just replace each occurence of "-----------" with its respective page number?

Would be nice if you could approve this @hexapode

[anoopshrma]

Hey @Laktus , json mode is out for quite some time now.
You can give it a try: https://github.com/run-llama/llama_parse/blob/main/examples/demo_json_parsing.ipynb

[Laktus]

@anoopshrma Hi, is there also LlamaIndexTS support? Can i somewhere see a list of all supported modes? (Markdown, JSON, any more?). Thanks.

[salujav4]

Hey @Laktus, I am facing the same issue, I am using LlamaParseReader in typescript and I don't see the json format there, just text and markdown

[BinaryBrain]

You can now retrieve page numbers and other information using the Json output.

Using the API:

GET https://api.cloud.llamaindex.ai/api/parsing/job/<jobId>/result/json

Using the TS lib:

const reader = new LlamaParseReader({ resultType: "json" });
const json = await reader.loadJson(path);

[thisadev]

Hi, Can we get the page number using Python?

[aceci0127]

You can now retrieve page numbers and other information using the Json output.

Using the API:

GET https://api.cloud.llamaindex.ai/api/parsing/job/<jobId>/result/json

Using the TS lib:

const reader = new LlamaParseReader({ resultType: "json" });
const json = await reader.loadJson(path);

what is the JobID?

[BinaryBrain]

Hi, Can we get the page number using Python?

Yes, if you get the Json results.

You can also use {pageNumber} in the page separator and page prefix and suffix

what is the JobID?

https://docs.cloud.llamaindex.ai/llamaparse/getting_started/api
Using the API, you first upload your file using https://api.cloud.llamaindex.ai/api/parsing/upload endpoint.
It'll return a jobID, and with the jobId, you can retrieve results.