kurwjan/lanisapi

An unofficial python library for Lanis.

Cześć!

I am 𝚔𝚞𝚛𝚠𝚓𝚊𝚗, a rookie developer from 𝙿𝚘𝚕𝚊𝚗𝚍 in 𝙶𝚎𝚛𝚖𝚊𝚗𝚢 with some experience. I mostly just experimented with languages and didn't do any 𝚋𝚒𝚐 projects until now.

You can contact me through:

• 𝙳𝚒𝚜𝚌𝚘𝚛𝚍: @kurwjan
• 𝙴-𝙼𝚊𝚒𝚕: [email protected]

Working on

• 𝙻𝚊𝚗𝚒𝚜𝙰𝙿𝙸, a python library for Hesse's Schulportal created by 𝚖𝚎.
• 𝚂𝙿𝙷-𝚟𝚎𝚛𝚝𝚛𝚎𝚝𝚞𝚗𝚐𝚜𝚙𝚕𝚊𝚗, a flutter app for Hesse's Schulportal created by @𝚊𝚕𝚎𝚜𝚜𝚒𝚘𝙲𝟺𝟸.

My skills 🔧

Good

Mediocre

Can exit

Stats 📈

lanisapi's Issues

⚡ - Add course class

This contains current tasks, past tasks, teacher, when, room, presence and so forth.

Calendar Feature Idea

Es wäre nützlich, wenn bei Kalendereinträgen auch steht von welchem Lehrer dieser erstellt wurde.
Dafür muss man den Wert von "Verantwortlich", der aus dem Json von Calendar kommt, mit der Lehrer ID verknüpfen.
An die IDs der Lehrer kommt man indem man alle Variationen von aa bis zz durchgeht. Dies wird bei der URL "https://start.schulportal.hessen.de/nachrichten.php?q={search}&page=1&a=searchRecipt" gemacht und gespeichert.

substitution error

Wenn die Schule auf Lanis keinen Vertretungsplan hat gibt es einen "list index out of range" error.
Vielleicht eine Prüfung auf Existenz einbauen. (Ich kann, wenn es gebraucht wird, auch die HTML schicken.)

File "***\Lib\site-packages\lanisapi\functions\substitution.py", line 84, in _get_substitution_info
date = re.findall(r"(\d\d.\d\d.\d\d\d\d)", html.html)[0]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^
IndexError: list index out of range

No Substitution error ?

I ran the example code with my credentials and ended with this:

INFO - LanisClient   Authenticated.
INFO - httpx   HTTP Request: POST https://start.schulportal.hessen.de/vertretungsplan.php "HTTP/1.1 200 OK"
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
File /opt/conda/lib/python3.11/site-packages/lanisapi/functions/substitution.py:88, in _get_substitution_info()
     87 try:
---> 88     date = re.findall(r"(\d\d\.\d\d\.\d\d\d\d)", date_element.text())[0]
     89 except AttributeError as err:

AttributeError: 'NoneType' object has no attribute 'text'

...

File /opt/conda/lib/python3.11/site-packages/lanisapi/functions/substitution.py:94, in _get_substitution_info()
     90     HTMLLogger.log_missing_element(
     91         html.html, "get_substitution_info()", "/", "date"
     92     )
     93     msg = "Critical date element was not found, something is definitely wrong! Please file a bug with the html_logs.txt file."
---> 94     raise CriticalElementWasNotFoundError(msg) from err
     96 LOGGER.info(f"Substitution info: Successfully got info. Notice is {bool(notice)}.")
     98 return {"notice": notice, "date": date}

CriticalElementWasNotFoundError: Critical date element was not found, something is definitely wrong! Please file a bug with the html_logs.txt file.

This is the actual conent of the page:

I can see the text in the html_logs.txt

html_logs.txt

Let me know what help can look like

⚡ - Add fetching "Lerngruppen" page

This page has all courses, teachers (also has profile pics of them) and tests.
Of course it has no API,, thanks Lanis ):

⚡ - Add fetching user details

• Get user details
• Get current profile picture

🔧 - Switch from dataclasses to attrs

Use attrs (or pydantic?) instead of Pythons dataclasses.
You can then also use its .asdict() to get JSON-like data.

• Replace dataclasses with attrs
• Return post link when you use asdict() on Calendar

zusätzlicher Tag

LanisAPI/src/lanisapi/functions/substitution.py

Line 129 in e39807b

lädt einen zusätzlichen Tag ... (unsere Schule hat immer 2 Tage einmal den aktuellen, dazu noch den nächsten Tag)

    # Add 1 day to the date
    tomorrow = date_obj + timedelta(days=1)

    # Create the list with the date and tomorrow
    date_list = [info["date"], tomorrow.strftime("%d.%m.%Y")]

    for date_str in date_list:
        print(f"Date: {date_str}")

        data = {"ganzerPlan": "true", "tag": date_str}

⚡ - Add fetching "Videokonferenz" page

It is probably used by no one anymore but it still exists to be parsed!

• Fetch page

Seems to be that no API exists.

⚡ - Add fetching "Stundenplan" page

• Get whole plan
• Get personal plan
• Get whole plan + possible substitutions (For some schools this is already implemented in Lanis but not for my school)

Looks like an API exists for this page.

📚 - Switch from github wiki to readthedocs

Is it unnecessary? Maybe but I also want to learn new stuff with this project so why not.

⚡ - When authenticating look for all supported lanis features.

🐍 - Some issues with retrieving saved sessions.

For example: sometimes I get a LoginPageRedirectError when using this lib normally.

• Fix LoginPageRedirectError

⚡ - Add option to use (and get) school name and location to authenticate.

Instead of

LanisClient("schoolid", ....)

You can use school name and location

LanisClient("schoolname", "location")

When you authenticate for the first time it parses this page to get all schools. (Probably in JSON format)

Also add a function to get all schools for the developer.

🔨 - Improve current logger

I mean the current one isn't so helpful.

Feste ID

In https://github.com/kurwjan/LanisAPI/blob/master/src/lanisapi/authentication_functions.py in Zeile 54 gibt es eine Feste School ID.

Rewrite

Hab wieder Bock mal auf Python und hab auch paar Ideen die dieses Projekt verbessern würde, aber müsste somit auch die komplette Struktur verändern. (villt. unnötig, aber egal)

Das wird wahrscheinlich nicht so schnell gehen, weil ich möchte auch nebenbei bei der Lanis-App weitermachen.

Mein Ziel ist es die gleichen Apps wie vorher zu unterstützen und den Rest kann man einfach selber erweitern.

Außerdem will ich Readme und Docs nicht mehr krankhaft in Englisch machen sondern einfach Deutsch.

🐛 Sometimes get_conversations() failes on the first try.

The problem is that unread doesn't exist?
Seems to be really random.

_get_available_apps() bug

Es wird "Mein Unterricht" nicht gefunden, weil es in implemented_apps "mein Unterricht" geschrieben wurde, aber dort "Mein Unterricht" stehen muss. (Wenn es geändert wurde, wird MeinUntericht gefunden.)

Tasks error

Wenn es einen Task/Eintrag gibt ohne Titel gibt es einen Crash.

File "***\Lib\site-packages\lanisapi\functions\tasks.py", line 67, in _get_tasks
title = element.css_first("td:nth-child(2) > b:nth-child(1)").text()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'NoneType' object has no attribute 'text'

⚡ - Add fetching messages

This is a bit harder because the "Nachrichten" page has a 2-Level-Encryption.

⚡ - Add CLI support

Add CLI support using Typer.
Would be pretty useful for (bash) scripts or so.

"Mein Unterricht" feature Idea

Für jedes Unterrichtsfach gibt es auch eine Detaillierte Seite, die mehr Informationen besitzt. Diese Seite könnte man auch einbauen, um eine Historie anzuzeigen.
https://start.schulportal.hessen.de/meinunterricht.php?a=sus_view&id=1234

🔨 - Save cookies after close()

As stated here the session is for 100min valid after the last request.
So when we call close() we could save the session cookies (encrypted) and on the next authenticate() we just load it.
Also add support for 30 days long sessions.

30 days session

It uses the same URLs as the normal login but it additionally posts to https://start.schulportal.hessen.de/registerbrowser and saves the cookie SPH-AutoLogin.
What to post is unknown.

• Update close() function ed600cf
• Update authenticate() function ed600cf
• Refactor code
• Add support for 30 days sessions

Remove encryption because it's unnecessary.

🔨 - Improve get tasks function

• Fetch grades, tests, presence and history from task detail page. Suggested by @baum-eule in #21
• Rename get_tasks function to get_current_tasks and add option to get only unfinished ones
• Get all subjects from "Kursmappen"
• Get presence from all subjects
• Get link of each file not only the whole .zip file.
• Set a task as finished or unfinished Will be moved to another milestone, grouped together with other POST actions.

🐎 - Switch from "normal" return to yield

Use Python generators for more performance and seems to be the preferred way in web scraping (see Scrapy).

📚 - Add help for scraping Lanis

Add to the docs handy CSS selectors, links or something similar for every thing what you can scrape.

• Add the handy things to the docs.

💬 - Add some comments to the code