Move scripts into own folder

apiserver/scripts/delete-story.py

@@ -0,0 +1,50 @@
+import database
+import search
+import sys
+
+import json
+import requests
+
+database.init()
+search.init()
+
+def database_del_story(sid):
+    try:
+        session = database.Session()
+        session.query(database.Story).filter(database.Story.sid==sid).delete()
+        session.commit()
+    except:
+        session.rollback()
+        raise
+    finally:
+        session.close()
+
+def search_del_story(sid):
+    try:
+        r = requests.delete(search.MEILI_URL + 'indexes/qotnews/documents/'+sid, timeout=2)
+        if r.status_code != 202:
+            raise Exception('Bad response code ' + str(r.status_code))
+        return r.json()
+    except KeyboardInterrupt:
+        raise
+    except BaseException as e:
+        logging.error('Problem deleting MeiliSearch story: {}'.format(str(e)))
+        return False
+
+if __name__ == '__main__':
+    if len(sys.argv) == 2:
+        sid = sys.argv[1]
+    else:
+        print('Usage: python delete-story.py [story id]')
+        exit(1)
+
+    story = database.get_story(sid)
+
+    if story:
+        print('Deleting story:')
+        print(story.title)
+        database_del_story(sid)
+        search_del_story(sid)
+        database.del_ref(story.ref)
+    else:
+        print('Story not found. Exiting.')

apiserver/scripts/fix-stories.py

@@ -0,0 +1,58 @@
+import time
+import json
+import logging
+
+import feed
+import database
+import search
+
+database.init()
+
+def fix_gzip_bug(story_list):
+    FIX_THRESHOLD = 150
+
+    count = 1
+    for sid in story_list:
+        try:
+            sid = sid[0]
+            story = database.get_story(sid)
+            full_json = json.loads(story.full_json)
+            meta_json = json.loads(story.meta_json)
+            text = full_json.get('text', '')
+
+            count = text.count('<EFBFBD>')
+            if not count: continue
+
+            ratio = count / len(text) * 1000
+            print('Bad story:', sid, 'Num ?:', count, 'Ratio:', ratio)
+            if ratio < FIX_THRESHOLD: continue
+
+            print('Attempting to fix...')
+
+            valid = feed.update_story(meta_json, is_manual=True)
+            if valid:
+                database.put_story(meta_json)
+                search.put_story(meta_json)
+                print('Success')
+            else:
+                print('Story was not valid')
+
+            time.sleep(3)
+
+        except KeyboardInterrupt:
+            raise
+        except BaseException as e:
+            logging.exception(e)
+            breakpoint()
+
+if __name__ == '__main__':
+    num_stories = database.count_stories()
+
+    print('Fix {} stories?'.format(num_stories))
+    print('Press ENTER to continue, ctrl-c to cancel')
+    input()
+
+    story_list = database.get_story_list()
+
+    fix_gzip_bug(story_list)
+

apiserver/scripts/reindex.py

@@ -0,0 +1,62 @@
+import logging
+logging.basicConfig(
+        format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+        level=logging.INFO)
+
+import database
+from sqlalchemy import select
+import search
+import sys
+
+import time
+import json
+import requests
+
+database.init()
+search.init()
+
+BATCH_SIZE = 5000
+
+def put_stories(stories):
+    return search.meili_api(requests.post, 'indexes/qotnews/documents', stories)
+
+def get_update(update_id):
+    return search.meili_api(requests.get, 'tasks/{}'.format(update_id))
+
+if __name__ == '__main__':
+    num_stories = database.count_stories()
+
+    print('Reindex {} stories?'.format(num_stories))
+    print('Press ENTER to continue, ctrl-c to cancel')
+    input()
+
+    story_list = database.get_story_list()
+
+    count = 1
+    while len(story_list):
+        stories = []
+
+        for _ in range(BATCH_SIZE):
+            try:
+                sid = story_list.pop()
+            except IndexError:
+                break
+
+            story = database.get_story(sid)
+            print('Indexing {}/{} id: {} title: {}'.format(count, num_stories, sid[0], story.title))
+            story_obj = json.loads(story.meta_json)
+            stories.append(story_obj)
+            count += 1
+
+        res = put_stories(stories)
+        update_id = res['uid']
+
+        print('Waiting for processing', end='')
+        while get_update(update_id)['status'] != 'succeeded':
+            time.sleep(0.5)
+            print('.', end='', flush=True)
+
+        print()
+
+    print('Done.')
+

apiserver/scripts/tests.py

@@ -0,0 +1,23 @@
+import time
+import requests
+
+def test_search_api():
+    num_tests = 100
+    total_time = 0
+
+    for i in range(num_tests):
+        start = time.time()
+
+        res = requests.get('http://127.0.0.1:33842/api/search?q=iphone')
+        res.raise_for_status()
+
+        duration = time.time() - start
+        total_time += duration
+
+    avg_time = total_time / num_tests
+
+    print('Average search time:', avg_time)
+
+
+if __name__ == '__main__':
+    test_search_api()