Skip to content

Latest commit

 

History

History
79 lines (67 loc) · 3.55 KB

CHANGELOG.md

File metadata and controls

79 lines (67 loc) · 3.55 KB

stocksight Change Log

[0.1-b.12] = 2020-06-08

changed

  • removed --noelasticsearch cli arg option

[0.1-b.11] = 2020-05-24

changed

  • removed stocksight web site uploading capability, site has been removed

[0.1-b.10] = 2020-03-29

added

  • Dockerfile and docker-compose.yml for running in docker

changed

  • added random time delay between fetching tweets to reduce chance of getting Twitter 420 code (throttled/backoff)

fixed

  • using -k keywords no longer causes twitter user id's to be looked up
  • fatal error when looking up and using twitter user ids

[0.1-b.9] = 2019-10-27

added

  • -l --linksentiment cli arg - follow any tweet link urls and run sentiment analysis on those web pages
  • requirement for newspaper3k python module to requirements.txt, install with pip

[0.1-b.8] = 2019-10-25

added

  • -w --websentiment cli arg - Get sentiment results from text processing website
  • improved nltk token processing - no longed needed to provide multiple case in nltk tokens in config
  • requirement for nltk stopwords, install with python -c "import nltk; nltk.download('stopwords')"

changed

  • getting web sentiment results from text processing website is no longer default
  • improved tweet text cleaning, sentiment algorithm and stocksight sentiment upload values

[0.1-b.7] = 2019-10-24

added

  • check if running Python 3
  • -U --upload - uploads sentiment to stocksight website (BETA) https://stocksight.diskoverspace.com/
  • stocksight_token in config.py.sample, used for auth to upload to stocksight website, copy to your config
  • nltk_min_tokens in config.py.sample, used to set minimum number of tokens required, copy to your config
  • tweet/news headline count/filtered/ratio log output
  • --noelasticsearch cli arg for not adding new docs to Elasticsearch
  • -s stock symbol cli arg (required arg), this is the stock symbol name and also used as the "tag name" on the stocksight webiste when uploading sentiment data
  • --overridetokensreq and --overridetokensignore cli args
  • -a --addtokens cli arg to add nltk required tokens from config to keywords

changed

  • --newsheadlines no longer requires stock symbol, use -s to provide stock symbol
  • nltk required tokens from config now do not automatically get added to keywords, use -a or --addtokens to add them

fixed

  • 'NoneType' object is not iterable Can't get sentiment from url caused by 400 Form Validation Errors text: This field is required traceback error when tweet with no text passed to sentiment_analysis

[0.1-b.6] = 2019-07-15

fixed

  • "TypeError: sequence item 0: expected str instance, int found" traceback error when running with -f twitteruserids.txt

[0.1-b.5] = 2019-01-11

changed

  • set encoding to utf-8 and checked for bytes when writing to twitteruserids.txt

[0.1-b.4] = 2018-12-10

fixed

  • TypeError: can't concat str to bytes when writing to twitteruserids.txt

[0.1-b.3] = 2018-11-23

added

  • requirements.txt for installing python requirements with pip
  • config.py.sample has new setting for specifying elasticsearch host/ip, port, username and password, copy to your config file

[0.1-b.2] = 2018-10-10

added

  • cli option -n --newsheadlines to fetch and analyze stock symbol headlines from yahoo finance website instead of twitter
  • cli option --frequency to control how often news headlines are retrieved
  • cli option --followlinks to follow any links in news headlines and scrape any relevant text on landing page
  • additional mappings for newsheadline docs in elasticsearch indices

changed

  • code cleanup

[0.1-b.1] = 2018-10-09

note

  • first release