search

theycallmeloki / askfm-python

A python requests + lxml based client side library for Ask.fm
5 stars 3 forks source link

Store answer IDs #1

Open sean-duffy opened 9 years ago

sean-duffy commented 9 years ago

First of all great job with this library, it does a very good job of getting the data and putting it into a convenient structure. I do have a few small fixes which I'll submit pull requests for when I've pushed them to my fork.

On to the matter at hand, I noticed you don't store any kind of unique identifier or timestamp for the answers at the moment. The timestamps are tricky since all ask.fm gives us is an interval, not a timestamp for when the question is posted. However, there is an ID associated with each answer which I think increases with each question, so it can be used to put a user's answers into the correct order.

This ID is found in the id attribute of each questionBox div, for example: <div class="questionBox" id="question_box_127193741364">

Let me know what you think, and keep up the good work!

theycallmeloki commented 9 years ago

Thanks a ton about the appreciation,

I've never personally had any reason to put a user's answers into the correct order, in my head, as long as the answers are present, it's all well and good. But I do realize now how that might help, and I agree, I'll add the ID into each answer, in the immediate future.

I'm not sure if you've noticed something, but if you google any user, the search results points back to ask.fm/username/best. The /best sorts the user's answers based on how many likes it's gathered. My original notion was to pick up answers through this, that's kind of why I didn't bother with the id attribute.

The unicode support is not terribly high on my priority list, but I must admit I wish I had it sooner, when you start scraping non-english speaking users it begins to really show why it is terribly important to have unicode support. I honestly know gracefully little about encoding and would love to learn how to fix this. Right now, you're right, I just gracefully ignore unicode.

Thanks for your interest in the library :)