Cleanup for public

Author: Greg Bell

README.md

@@ -0,0 +1,28 @@
+# Bayesian RSS feed classifier
+Downloads news articles from an RSS feed, categorises them based on naive Bayesian classification, then prints to screen and emails.
+
+Only articles not seen before are shown.
+
+This project makes use of the awesome [toolz](https://github.com/pytoolz/toolz)
+library to organize everything as a pipeline of functions.
+
+## Suitability
+This is a personal project developed to process the Port News and Wauchope
+Gazette newspapers in Australia. XML/HTML tags and properties will be specific
+to those feeds, so you'll have to modify code for your particular feed. If
+there's interest, complain about this in an issue, and I'll make everything
+configurable.
+
+It also only classifies articles as "news" or "sports".
+
+## Training
+Use `download_article.py` to download a particular article, based on URL, and
+manually put the resulting file in either `training_data/news` or
+`training_data/sports` dir.
+
+## Other setup
+Add your feed URLs and email to `feed_classifier.py`
+Configure your email server in `pipeline/mail_sender.py`
+
+## Running
+Run `feed_classifier.py`, no arguments required.

app/feed_classifier.py

@@ -40,17 +40,14 @@
 # Dependencies injected for testing purposes
 # Note there's a 'dummy' module which has versions of the functions in
 # `internet` that don't connect to the internet. For end-to-end testing.
-# TODO:
-# implement dummy versions of the persistence functions, also for testing.
 from feed_classifier.internet import feed_fetcher, body_fetcher
 from feed_classifier.NB_classifier import NB_classifier
 from feed_classifier.persistence import have_seen
 
 
 def main():
+    # URL to rss.xml's
     URLS = [
-        "http://www.portnews.com.au/rss.xml",
-        "https://www.wauchopegazette.com.au/rss.xml",
     ]
 
     title_hashes = load_title_hashes()
@@ -74,7 +71,7 @@ def main():
             mark_as_seen(title_hashes),
             format_as_text,
             toolz.juxt(
-                print, sendmail(["greg@bitwombat.com.au", "jillbell2@yahoo.com"])
+                print, sendmail([])
             ),
         ),
     )

app/pipeline/mail_sender.py

@@ -6,8 +6,8 @@
 
 @autocurry
 def sendmail(recipients, text):
-    SERVER = "server"
-    FROM = "greg@gbell.bitwombat.com.au"
+    SERVER = ""
+    FROM = ""
     SUBJECT = "New Local Stories"
 
     email = """\