[cad161]: / docs / data / json.md

Download this file

44 lines (32 with data), 1.2 kB

JSON

??? abstract "TLDR"

```{ .python .no-check }
import edsnlp

stream = edsnlp.data.read_json(path, converter="omop")
stream = stream.map_pipeline(nlp)
res = stream.to_json(path, converter="omop")
# or equivalently
edsnlp.data.to_json(stream, path, converter="omop")
```

We provide methods to read and write documents (raw or annotated) from and to json files.

As an example, imagine that we have the following document that uses the OMOP schema

{ "note_id": 0, "note_text": "Le patient ...", "note_datetime": "2021-10-23", "entities": [...] }
{ "note_id": 1, "note_text": "Autre doc ...", "note_datetime": "2022-12-24", "entities": [] }
...

You could also have multiple .json files in a directory, the reader will read them all.

Reading JSON files {: #edsnlp.data.json.read_json }

::: edsnlp.data.json.read_json
options:
heading_level: 3
show_source: false
show_toc: false
show_bases: false

Writing JSON files {: #edsnlp.data.json.write_json }

::: edsnlp.data.json.write_json
options:
heading_level: 3
show_source: false
show_toc: false
show_bases: false