No examples in the documentation (HTTP API)

Rasa Open Source
1

I can’t understand the http://rasa.com/docs/rasa/api/http-api/#tag/Model/paths/~1model~1train/post in order to post my training data to my server. There is no example given. Previous versions at least had examples but now, the docs are a lot messy.

Can someone provide me with an example? Can maybe the Rasa developers take this into account? There are many users in the forum that are asking similar questions.

1 Like
2

Hi @eloukas, what exactly do you mean by example? There is an example request body in the HTTP API on the right. However, I just realised that we are missing some line breaks here. Will update that. What else are you missing? Your feedback is highly appreciated. Thanks!

In order to train a model via the API, you need to have a request body similar to this

{
    "config": "language: en\npipeline: supervised_embeddings\npolicies:\n  - name: MemoizationPolicy\n  - name: KerasPolicy",
    "domain": "intents:\n  - greet\n  - goodbye\n  - mood_affirm\n  - mood_deny\n  - mood_great\n  - mood_unhappy\nactions:\n  - utter_greet\n  - utter_cheer_up\n  - utter_did_that_help\n  - utter_happy\n  - utter_goodbye\ntemplates:\n  utter_greet:\n  - text: \"Hey! How are you?\"\n  utter_cheer_up:\n  - text: \"Here is something to cheer you up:\"\n    image: \"https://i.imgur.com/nGF1K8f.jpg\"\n  utter_did_that_help:\n  - text: \"Did that help you?\"\n  utter_happy:\n  - text: \"Great carry on!\"\n  utter_goodbye:\n  - text: \"Bye\"",
    "nlu": "## intent:greet\n  - hey\n  - hello\n## intent:goodbye\n  - cu\n  - good bye",
    "stories": "## happy path\n* greet\n  - utter_greet\n* mood_great\n  - utter_happy\n## sad path\n* greet\n  - utter_greet\n* mood_unhappy",
    "out": "models",
    "force": false
}

To actually train a model you can use

curl -X POST \
  http://localhost:5005/model/train \
  -H 'Content-Type: application/json' \
  -H 'accept: application/octet-stream' \
  -d '{
    "config": "language: en\npipeline: supervised_embeddings\npolicies:\n  - name: MemoizationPolicy\n  - name: KerasPolicy",
    "domain": "intents:\n  - greet\n  - goodbye\n  - mood_affirm\n  - mood_deny\n  - mood_great\n  - mood_unhappy\nactions:\n  - utter_greet\n  - utter_cheer_up\n  - utter_did_that_help\n  - utter_happy\n  - utter_goodbye\ntemplates:\n  utter_greet:\n  - text: \"Hey! How are you?\"\n  utter_cheer_up:\n  - text: \"Here is something to cheer you up:\"\n    image: \"https://i.imgur.com/nGF1K8f.jpg\"\n  utter_did_that_help:\n  - text: \"Did that help you?\"\n  utter_happy:\n  - text: \"Great carry on!\"\n  utter_goodbye:\n  - text: \"Bye\"",
    "nlu": "## intent:greet\n  - hey\n  - hello\n## intent:goodbye\n  - cu\n  - good bye",
    "stories": "## happy path\n* greet\n  - utter_greet\n* mood_great\n  - utter_happy\n## sad path\n* greet\n  - utter_greet\n* mood_unhappy",
    "out": "models",
    "force": false
}'

Does this help?

2 Likes
3

Hi @Tanja i am using curl -X POST http://localhost:5005/conversations/user1/messages -d '{"sender": "user1", "text": "hello"}' to add the message to the tracker but i am facing an issue as {"version":"1.10.3","status":"failure","message":"Currently, only user messages can be passed to this endpoint. Messages of sender 'user1' cannot be handled.","reason":"BadRequest","details":{"parameter":"sender","in":"body"},"help":null,"code":400}

If you can help. Thanks

4

Can you try to pass in "sender": "user" instead of "sender": "user1"? Everything else should stay the same.