Initial commit

.gitignore

@@ -0,0 +1,5 @@
+*.swp
+.idea
+*.pyc
+data/
+venv/

README.md

@@ -0,0 +1,9 @@
+Download and format Google Play stats as single JSON file
+
+## Requirements
+
+You need [gsutil](https://cloud.google.com/storage/docs/gsutil_install)
+
+## Usage
+
+    python google_play.py ACCOUNT_ID

google_play.py

@@ -0,0 +1,53 @@
+import json
+import subprocess
+
+import os
+from csv import DictReader
+
+from collections import defaultdict
+
+import sys
+
+DATA_DIR = 'data'
+
+
+def download_csvs(account_id):
+    if not os.path.exists(DATA_DIR):
+        os.mkdir(DATA_DIR)
+
+    uri = 'gs://pubsite__rev_{}/stats/installs/*_overview.csv'.format(account_id)
+    cmd = ['gsutil', 'cp', '-r', uri, 'data']
+    p = subprocess.Popen(cmd, stderr=subprocess.PIPE, stdout=subprocess.PIPE)
+    out, err = p.communicate()
+    if p.returncode != 0:
+        print(err.decode('utf-8'))
+        exit(2)
+
+
+def to_json():
+    stats = defaultdict(list)
+    for dirpath, dnames, file_names in os.walk(DATA_DIR):
+        for file_name in file_names:
+            path = os.path.join(dirpath, file_name)
+            if path.endswith('.csv'):
+                with open(path, encoding='utf-16') as fp:
+                    for line in DictReader(fp):
+                        stats[line['Package Name']].append(line)
+
+    # Sort
+    for pkg, lines in stats.items():
+        stats[pkg] = sorted(lines, key=lambda x: x['Date'])
+
+    # Write
+    for pkg, lines in stats.items():
+        with open(os.path.join(DATA_DIR, '{}.json'.format(pkg)), 'w', encoding='utf-8') as fp:
+            json.dump(lines, fp, indent=4)
+
+
+if __name__ == '__main__':
+    if len(sys.argv) != 2:
+        print("Usage: python google_play.py ACCOUNT_ID")
+        exit(1)
+
+    download_csvs(sys.argv[1])
+    to_json()