tp ush

Merge branch 'master' of github.com:saasify-sh/twitter-flock

* 'master' of github.com:saasify-sh/twitter-flock:
  Add ranking options by followers, age, verification (#5)

lib/ranking.js

@@ -0,0 +1,23 @@
+'use strict'
+
+// TODO: Serialization mechanism to pass data objects.
+// Options:
+// - List of user id's in stdout
+// - JSON in stdout
+// - JSON in output file
+// - SQLite database
+
+const spawn = require('child_process').spawn
+const pythonProcess = spawn('python3', ['ranking/rank.py'])
+
+pythonProcess.stdout.on('data', (data) => {
+  // Do something with the data returned from python script
+  console.log('received data from python stdout')
+  console.log(data.toString())
+})
+
+pythonProcess.stderr.on('data', (data) => {
+  // Do something with the data returned from python script
+  console.log('received data from python stderr')
+  console.log(data.toString())
+})

ranking/rank.py

@@ -0,0 +1,47 @@
+import pandas as pd
+import json
+import pprint
+
+# These methods are all functional
+def filter_verified(df) -> pd.DataFrame:
+    return df[df['verified'] == True]
+
+def rank_by_col(df, col: str, ascending: bool) -> pd.DataFrame:
+    return df.sort_values(col, ascending=ascending)
+
+def rank_by_lambda(df, lmbda, ascending: bool) -> pd.DataFrame:
+    """ Lambda function should have a single DataFrame argument and return a Series.
+
+    For example, `lmbda = lambda df: df.followers_count * df.friends_count`.
+    """
+    return df.loc[lmbda(df).sort_values(ascending=ascending).index]
+
+if __name__ == "__main__":
+    print("Running 'python ranking/test.py'")
+    pd.set_option('display.max_columns', None)
+
+    with open("fixtures/twitter-users-10k.json", "r") as read_file:
+        users: list = json.load(read_file)
+
+    cols = ["name", "screen_name", "description", "url", "followers_count", "friends_count",
+            "favourites_count",  "created_at", "verified", "following", "status", "id", "id_str"]
+    df = pd.DataFrame(users)[cols]
+    df["created_at"] = pd.to_datetime(df["created_at"])
+
+    print(df)
+    # Filter by verified
+    print("Filter verified")
+    print(filter_verified(df))
+    # Filter by followers
+    print("By followers")
+    print(rank_by_col(df, 'followers_count', ascending=False))
+    # Filter by friends (someone who follows you and nobody else is likely a superfan)
+    print("By friends")
+    print(rank_by_col(df, 'friends_count', ascending=True))
+    # Filter by favourites activity
+    print("By favourites")
+    print(rank_by_col(df, 'favourites_count', ascending=False))
+    # Filter by follower ratio
+    rank_by_lambda(df, lambda df: df.followers_count - df.friends_count, ascending=False)
+    # Filter by account age
+    print(rank_by_col(df, 'created_at', ascending=True))