tweet-model-serpucga 0.7.3

A modelization of a tweet object with convenient features and functionalities

Tweet Model

A modelization of a tweet object with convenient features and functionalities. This project was thought for integration and usage as a package by the Twitter Dashboard project.

• Free software: MIT license

• Documentation: https://tweet-model.readthedocs.io.

Features

• A modelization of a tweet in the form of class Tweet. This class contains a constructor that initializes all the possible tweet attributes to None except those indicated otherwise.

• The inner objects of a tweet (“user”, “entities”, “places”, etc.) are stored internally as nested dictionaries.

• The __getitem__() method for Tweet is overriden to allow a dictionary-like access to the tweet contents. For example, if “tweet1” is an instance of Tweet, one could do tweet1[“id”] to get the id of that tweet, or tweet1[“user”][“name”] to get the name of the person that published the tweet.

Credits

Creator: Sergio

This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.

History

0.1.0 (2019-03-13)

• First release on PyPI.

0.2.0 (2019-03-20)

• Completed the Tweet class that allows the user to make usable instances of a tweet model. Includes initialization of all the Tweet attributes indicated in the Twitter documentation (default to None, unless the user provides a value) and overriding of __getitem__ to provide a dictionary-like access to the information.

0.3.0 (2019-03-20)

• Added method “get_tweets_from_csv()”, which gets a CSV file as an argument and returns a list containing as many Tweet objects as lines (minus the header) in the CSV file. The header of the CSV is used to know which attributes should be set.

• The method will raise an error and exit if any item in the header does not match with the specification of the Tweet object (for example, the header word “media.sizes.thumb.h” would be valid, but “user.lightsaber.color” would not.

• At this point, the method took 1.75s aprox to read and return the contents of a 5.7 MB as a list of ‘Tweet’s. This could be troublesome with very large collections in a future if the progression of time was proportional with the file size (estimation would be 25 minutes for a 5 GB file)

0.4.0 (2019-03-26)

Since the last release the main improvements have been:

• Raise a custom exception instead of sysexiting when confronted with a wrong CSV file.

• Change prints for logger messages.

• Modularize the “get_tweets_from_csv” dividing it into 3 functions to provide more isolated functions. Now this function makes use of “get_tweet_from_csv_line”, which returns a single Tweet object when provided with two lists representing the contents of a CSV header and one of its entries. There is also a “raw” alternative of that function, which works with raw lines and converts them to lists to then call the “non-raw” version.

Version 0.4.1: The tweet-manager dependency, necessary for the core functions, has been added into production (in setup.py). Version 0.4.2: Seems like that wasn’t the right way to add a requirement. Tried now pip freezing to requirements.txt. Version 0.4.3: Updated tweetmanager dependency to current version (1.1.5) Version 0.4.4: Fixed another requirement error Version 0.4.5: Changed f-strings to traditional strings for reasons of compatibility Version 0.4.6: Trying to track an error were Tweets are not instantiated if the CSV registry contains certain symbols Version 0.4.7: Added function to generate dicts representing subsets of Tweet content (return just the fields indicated by the user and not the full Tweet object)

0.5.0 (2019-05-03)

• Code from tweet_model.py now divided in two source files: tweet.py (containing the Tweet class) and utils.py (which contains the rest of the utilities to instantiate Tweet objects from CSVs, return a mini-dict with partial information of a Tweet, etc.)

• Added two new fields to the Tweet class: “polarity” and “trtext”, which are not part of the original Tweeter tweets, but needed in other project.

• Added setter methods for “polarity” and “trtext”, which will be probably now set at instantiation time but afterwards

Version 0.5.1: Tiny fix in the imports Version 0.5.2: Added method to get Tweet in JSON form Version 0.5.3: Coordinates stored as list and not as str Version 0.5.4: Minor fix Version 0.5.5: Implementing typings

0.6.0 (2019-05-16)

Main differences with last major release:

• Tweet objects do not store all the values as str anymore. The most used fields are now attempted to cast to their type on initialization. These improvements were added on the last versions of 0.5.x

• Tests have been added that cover most of the code in tweets.py and utils.py. In the next versions probably even more will be added.

Version 0.6.1: Added more tests and the “as_json()” function divided in two: “as_long_json”, which simply returns the __dict__ of the Tweet instance, with all its key-value pairs, and “as_short_json”, which returns just the non-empty key-value pairs (and it won’t even return nested dicts it they have no content). Version 0.6.2: No changes, just for trying new hooks. Version 0.6.3 & 0.6.4: Updated tweetmanager dependency.

0.7.0 (2019-10-14)

Main differences with last major release:

• Accept embedded tweets in “quoted_status” and “retweeted_status”.

• Accept some other new fields found in some tweets (although they don’t seem to appear in the Twitter documentation) such as “full_text” or “display_text_range”

• Accept any other argument without throwing an error, but don’t do shit with them. Thus, if unexpected and undocumented fields arrive, they won’t be used to instantiate the Tweet object, but it won’t crash either.

• Some reformatting

Version 0.7.1: Reworks on dependencies Version 0.7.2: Rework to accept “nan” on some fields and translate it as None Version 0.7.3: Accept any 1.1.x tweetmanager version instead of just 1.1.7