IGBloks

If you are currently also working on any form of implementation, researches, or any other thing about bloks, contact me on discord to be added in a private groupchat to share knowledges.

🧱 A library to work with instagram bloks technology

See the KNOWLEDGES for further explaination on what this is about.

⚠️ This library is currently being develloped, it is unstable.

Installation 🔌

# Easy install
pip install git+https://github.com/novitae/igbloks.git

# If you want to also access to the tools
git clone https://github.com/novitae/igbloks
cd igbloks
pip install .

Usage

BeautifulSoup 🍜

If you have read the knowledges paper, you know that bloks looks like HTML, but under JSON format. So this library includes a converter BLOKS to HTML. You can then use the result and parse it using BeautifulSoup. Here is an example:

Let's say I want to get the value from this blok tree, "May 2021": Which actually looks like this converted to HTML:

>>> from igbloks.html import bloks_to_html
>>> import json
>>>
>>> with open("bloks.json", "r") as read: # Opening my json file containing my blok response.
...     data = json.load(read)["layout"]["bloks_payload"]["tree"] # Placing the tree in variable `data`.
... 
>>> htmled = bloks_to_html(tree) # Turning the tree into an HTML string.
>>> # Here, `htmled` contains the same data as the screenshot upper.
>>> soup = BeautifulSoup() # Loading the HTMLed bloks into BeautifulSoup
>>>
>>> # Then, I will search for the attribute of the item that has "Date joined". Since it doesn't change
>>> # between the responses, it can be our stable point from which we can then search for the date that
>>> # changes depending of the accounts.
>>> dj = soup.find("bk.components.text", {"text": "Date joined"})
>>> # Make sure that the name is lowercased (`"bk.components.Text"` -> `"bk.components.text"`)
>>>
>>> # And then I get the parent of it, search for all the `"bk.components.text"` items, at take the last
>>> # one, which contains the date data.
>>> dj.parent.find_all("bk.components.text")[-1]["text"]
'May 2012'

In the case the bloks syntax isn't exportable to HTML and then readable correctly (with some spaces or chinese characters as keys), you can set as_hex on True when using bloks_to_html to replace the key by the value f"x{key.encode().hex().upper()}".

Mapping 🗺️

Due to the way instagram bloks are working, with keys changing based on the bkid, the mapping tool is here to identify objects with the different keys the might be identified by. It works using a set of responses to same requests, but with different bkids. It produces a file that the library will take as a reference to know which object has which values, their types, their required values, their names, keys, etc ... .

Steps:

1. Create a directory (you will need it since the program iterates through each sub directories of it). It will be our list of responses set directory.
2. Capture some blok requests with an http proxy. The example is request is the one you get from the "About this account" page on iOS. Here I have one from my iOS device, we can see the fields name being very small.
3. Save the json response body in a new directory inside the created directory, name it whatever you want. For the example, I will name mine with the appid of the blok request. Name the request with the bkid used to send it. For me, it is ae6e4040481b1c3ecd691a99c8cd3849d4a4b4b05259fb9ae75c92c4b6a722ca, as you can see on the upper screenshot.
4. Now, replay the request with a different bkid. You can change only the one in the body, the one in the headers (if there is) is taken only if the one in the body is missing. To find new bkid, you can check the requests of the web instagram, or from another device / version. Here I will use 456a06cb008a2f6d72f04e94d26242d4607e80d5cf8d9a5aa34ec109117071c6, which is a web one. Sometimes, certain requests with different bkid are returning response 404, with the message "Payload returned is null. A server error field_exception occured. Check server logs for details.". It means that the appid doesn't support the given bkid, we can't do anything about it ... . With the different bkid, you can see the names of fields are completely different from the first response, but the composition stay the same, with the same values.
5. Save this new response body to the same subdirectory you made, and name the json file with the bkid used.

• You can repeat the steps 4 and 5 as many times you want with different bkid.