abraham

PyPI PyPI - Downloads GitHub PyPI - Python Version GitHub issues GitHub last commit

Algorithmically predict public sentiment on a topic using flair sentiment analysis.

Installation

Installation is simple; just install via pip.

$ pip3 install abraham3k

Basic Usage

You can run the main function, news_sentiment to get the raw scores. This will return a nested dictionary with keys for each topic.

from abraham3k.prophets import Isaiah

darthvader = Isaiah(news_source="google") 

scores = darthvader.news_sentiment(["amd", 
                               "microsoft", 
                               "tesla", 
                               "theranos"], 
                               window=2)
print(scores['tesla']['text'])

'''
                                                 desc              datetime  probability sentiment
0   The latest PassMark ranking show AMD Intel swi...  2021-04-22T18:45:03Z     0.999276  NEGATIVE
1   The X570 chipset AMD offer advanced feature se...  2021-04-22T14:33:07Z     0.999649  POSITIVE
2   Apple released first developer beta macOS 11.4...  2021-04-21T19:10:02Z     0.990774  POSITIVE
3   Prepare terror PC. The release highly anticipa...  2021-04-22T18:00:02Z     0.839055  POSITIVE
4   Stressing ex x86 Canadian AI chip startup Tens...  2021-04-22T13:00:07Z     0.759295  POSITIVE
..                                                ...                   ...          ...       ...
95  Orthopaedic Medical Group Tampa Bay (OMG) exci...  2021-04-21T22:46:00Z     0.979155  POSITIVE
96  OtterBox appointed Leader, proudly 100% Austra...  2021-04-21T23:00:00Z     0.992927  POSITIVE
97  WATG, world's leading global destination hospi...  2021-04-21T22:52:00Z     0.993889  POSITIVE
98  AINQA Health Pte. Ltd. (Headquartered Singapor...  2021-04-22T02:30:00Z     0.641172  POSITIVE
99  Press Release Nokia publish first-quarter repo...  2021-04-22T05:00:00Z     0.894449  NEGATIVE
'''

Changing News Sources

Isaiah supports two news sources: Google News and NewsAPI. Default is Google News, but you can change it to NewsAPI by passing Isaiah(news_source='newsapi', api_key='<your api key') when instantiating. I'd highly recommend using NewsAPI. It's much better than the Google News API. Setup is really simple, just head to the register page and sign up to get your API key.

Detailed Usage

Currently, there are a couple extra options you can use to tweak the output.

When instatiating the class, you can pass up to five optional keyword arguments: news_source and api_key (as explained above), splitting, and weights.

loud: bool - Whether or not the classifier prints out each individual average or not. Default: False.
splitting: bool - Recursively splits a large text into sentences and analyzes each sentence individually, rather than examining the article as a block. Default: False.
weights: dict - This chooses what each individual category (text, title, desc) is weighted as (must add up to 1). Default: weights={"title": 0.1, "desc": 0.1, "text": 0.8}.

When running the main functions, news_sentiment and news_sentiment_summary, there is one requred argument, topics, and two optional keyword arguments: window and up_to.

topics: list - The list of the topics (each a str) to search for.
up_to: str - The latest day to search for, in ISO format (%Y-%m-%dT%H:%M:%SZ). Default: current date.
window: int - How many days back from up_to to search for. Default 2.

Updates

I've made it pretty simple (at least for me) to push updates. Once I'm in the directory, I can run $ ./build-push 1.2.0 "update install requirements" where 1.2.0 is the version and "update install requirements" is the git commit message. It will update to PyPi and to the github repository.

Notes

Currently, there's another algorithm in progress (SALT), including salt.py and salt.ipynb in the abraham3k/ directory and the entire models/ directory. They're not ready for use yet, so don't worry about importing them or anything.

Contributions

Pull requests welcome!

License
- OSI Approved :: MIT License
Operating System
- OS Independent
Programming Language
- Python :: 3

Release files for abraham3k 1.3.4

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for abraham3k 1.3.4
File	Size	Uploaded
abraham3k-1.3.4.tar.gz	11.6 MB	Apr 23, 2021	Details

Built distribution (wheel)

Table of built distributions (wheels) for abraham3k 1.3.4
File	Interpreter	ABI	Platform	Reset
abraham3k-1.3.4-py3-none-any.whl 26.4 kB Apr 23, 2021	Python 3	none	any	Details

Total release size: 11.7 MB

Release files / abraham3k-1.3.4.tar.gz

Download URL	abraham3k-1.3.4.tar.gz
Size	11.6 MB
Tags	Source
SHA-256 checksum How to use checksums	`c0ab34e54b0f51c4347690568f56589c901868b93418c3b14458e2586a73068c`
BLAKE2b-256 checksum How to use checksums	`4196789ad376ce28d4c604aa97883ce8129c0e7cd3cd01af03f04d7a55c04199`
Upload date	Apr 23, 2021
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/3.4.1 importlib_metadata/3.10.1 pkginfo/1.7.0 requests/2.25.1 requests-toolbelt/0.9.1 tqdm/4.58.0 CPython/3.9.2`

Release files / abraham3k-1.3.4-py3-none-any.whl

Download URL	abraham3k-1.3.4-py3-none-any.whl
Size	26.4 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`74104e056055767a04d9a0a3de24c12165e6c03e0e769bdb0f2886e081402808`
BLAKE2b-256 checksum How to use checksums	`23cb6434b5306f1476c87d2511addf5f294166339248e1247ca27fbdf0a7b470`
Upload date	Apr 23, 2021
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/3.4.1 importlib_metadata/3.10.1 pkginfo/1.7.0 requests/2.25.1 requests-toolbelt/0.9.1 tqdm/4.58.0 CPython/3.9.2`