Creation of a project: NER with relation
From the header, click on Create new project.
1/ Project tab
Fill in the required fields:
- Customer
- Project type
- Project name
- Project deadline
- Project description
Non-mandatory fields:
- Default markers: appears on the item later
- to delete: indicates an item to delete
- to check: indicates an item to check
- to discuss: indicates an item to discuss
You can define access levels for the imported item :
- the admin
- the data scientist
- the user
2/ Files tab
the fields :
- Config file: config file
- Annotations file: file containing the annotations linked to the item
- Predictions file: file containing the predictions linked to the item
Unlike annotations, predictions come from a pre-trained model. After the import, we can choose whether to keep them or not.
As for the mandatory field:
- item file: this is a list that must be prepared in advance including the type of item and a unique id
3/ Labeling tab
You can define what you want to do with the project.
- Task Title: defines the name
- Task Type: defines the type
You can add a new label by clicking on Add a new label.
The fields :
- Label Code: must be unique
- Label Name: does not necessarily have to have the same name as Label Code
- Color: defines the color of the label
- Hotkey: allows you to create a keyboard shortcut
- Label Description: describes the label
caution
Once these labels have been defined, it is no longer possible to modify them later.
4/ Annotation Guide tab
Here you can define annotation rules.
Once the specifications of a project have been filled in, you can click on Add a new task to start filling in the details of another project.
ex config.json
{
"tasks": [],
"name": "DEMO: NER with relations",
"client": "LJN",
"type": "text",
"highlights": [],
"description": "Projet de demo",
"admins": ["admin@test.com"],
"users": ["user@test.com"],
"dataScientists": ["data@test.com"],
"defaultTags": [],
"showPredictions": true,
"prefillPredictions": true,
"filterPredictionsMinimum": 0.4,
"deadline": "2023-11-30T13:57:20.355Z",
"entitiesRelationsGroup": [
{
"_id": "62de9a245902f5001ce49c77",
"name": "relationsGroup1",
"min": 2,
"max": 2,
"values": [
{
"exposed": true,
"_id": "62de9a245902f5001ce49c78",
"value": "is_from",
"label": "is_from"
},
{
"exposed": true,
"_id": "62de9a245902f5001ce49c79",
"value": "belongs_to",
"label": "belongs_to"
}
]
},
{
"_id": "62de9a245902f5001ce49c7a",
"name": "relationsGroup2",
"values": [
{
"exposed": true,
"_id": "62de9a245902f5001ce49c7b",
"value": "is_unit",
"label": "is_unit"
}
]
}
]
}
ex items.jsonlines
{
"predictions": {
"raw": { "Entities": { "entities": [{ "value": "name", "start": 23, "end": 33 }] } },
"keys": [{ "value": "name", "start": 23, "end": 33 }]
},
"uuid": "1831440",
"data": { "text": "En tant que militaire, volontaire sous contrat" },
"type": "text",
"metadata": {},
"description": "",
"annotated": false,
"createdAt": 1658755620660,
"velocity": null,
"lastAnnotator": {},
"seenAt": "2022-11-02T11:59:42.501Z"
}