Version scalable parsing ideas

[denizaydemir]

Hi, mapping pathings like "layout.bloks_payload.tree.㜍.$.㐈. .0.㐈. .0.㐅. .1.㐈. .0.㐈. .2.㐈. .0.㐈.�.1.㚝.&" is not really feasible since it becomes very unpredictable of how the api returns responses for differing scenarios of a same request. For example, if you have 5+ response scenarios for a login request(login success, wrong password, wrong account, 2fa redirect, trusted device verification, etc...), you should create a pathing of factorial of 5. Even they contain some of the same data in each, their pathings change. A robust method will be parsing the instruction sets of the UI navigation or next candidate requests. This way you could get a platform and bloks version independent parsing technique. One example is searching a candidate request url "com.bloks.www.two_step_verification.method_picker" and pick its instructions. However, I didn't dive deep into getting UI elements such as labels etc. Mapping pathings method is easier to get UI elements.

[novitae]

I see what you mean. In the base of the tree, from my memories, it is always the same structure (as long as you translate the keys for you block id), but then it gets sketchy. However, objects have the same architecture, and you can parse them like this.

For now I have set the project on the side, but I will come back to it in few times after I finish prior ones targeting instagram. I thought about translating the whole tree into an lxml object, that would be explorable using xpath. Lmk what are your thoughts on that.