chore(docs): add openAPI schema

[ruocco-l]

Closes #92

Adding openAPI for both RAG and URL2M

[nicklamonov]

As commented in Slack:

URL2M:

• the description of the endpoint says "An Apify Actor that takes....". Should we rather keep such actor description at the actor level and change the description for the endpoint a bit? Functionally it's fine.

RAG

• I also don't like a part of description. Maybe we should remove this - "Web browser for OpenAI Assistants API and RAG pipelines, similar to a web browser in ChatGPT."?
• At the same I would mention that if only the URL is provided, we do a fetch of that URL only. Now it's not explicit and humans/agents might not understand the difference.

[nicklamonov]

Thanks! All good from my side.

[nikitachapovskii-dev]

Thank you, @ruocco-l !

I noticed rag readme is updated for openapi but url2m not, maybe we should update it as well?

[nikitachapovskii-dev]

Did you skip some fields because they are hidden? (removeElementsCssSelector, desiredConcurrency, debugMode)
Should we in that case skip requestTimeoutSecs and debugMode for rag?

[ruocco-l]

Yeah we can do that

[nikitachapovskii-dev]

I asked Claude to compare this PR with the web-fetch one (apify/actor-web-fetch#1) and it flagged a difference that proxyConfiguration here has type: object while there they got headers type: string (JSON-encoded) for the GET
I guess we should align it here to be string as well so it JSON.parse's correclty

                    {
                        "name": "headers",
                        "in": "query",
                        "required": false,
                        "schema": {
                            "type": "string",
                            "description": "Additional HTTP headers to send with the request, as a JSON-encoded object string."
                        },
                        "example": "{\"Accept-Language\":\"de-DE\"}"
                    },

[ruocco-l]

Yeah, this is correct, I'll change it