this repo has no description
Documentation for follow/track commands, closes #11
+49
-1
+48
README.md
+48
README.md
···
146
146
--attach=foo:attendees.db \
147
147
--sql="select Twitter from foo.attendees"
148
148
149
+
## Capturing tweets in real-time with track and follow
150
+
151
+
This functionality is **experimental**. Please [file bug reports](https://github.com/dogsheep/twitter-to-sqlite/issues) if you find any!
152
+
153
+
Twitter provides a real-time API which can be used to subscribe to tweets as they happen. `twitter-to-sqlite` can use this API to continually update a SQLite database with tweets matching certain keywords, or referencing specific users.
154
+
155
+
### track
156
+
157
+
To track keywords, use the `track` command:
158
+
159
+
$ twitter-to-sqlite track tweets.db kakapo
160
+
161
+
This command will continue to run until you hit Ctrl+C. It will capture any tweets mentioning the keyword [kakap](https://en.wikipedia.org/wiki/Kakapo) and store them in the `tweets.db` database file.
162
+
163
+
You can pass multiple keywords as a space separated list. This will capture tweets matching either of those keywords.
164
+
165
+
$ twitter-to-sqlite track tweets.db kakapo raccoon
166
+
167
+
You can enclose phrases in quotes to search for tweets matching both of those keywords:
168
+
169
+
$ twitter-to-sqlite track tweets.db 'trash panda'
170
+
171
+
See [the Twitter track documentation](https://developer.twitter.com/en/docs/tweets/filter-realtime/guides/basic-stream-parameters#track) for advanced tips on using this command.
172
+
173
+
Add the `--verbose` option to see matching tweets (in their verbose JSON form) displayed to the terminal as they are captured:
174
+
175
+
$ twitter-to-sqlite track tweets.db raccoon --verbose
176
+
177
+
### follow
178
+
179
+
The `follow` command will capture all tweets that are relevant to one or more specific Twitter users.
180
+
181
+
$ twitter-to-sqlite follow tweets.db nytimes
182
+
183
+
This includes tweets by those users, tweets that reply to or quote those users and retweets by that user. See [the Twitter follow documentation](https://developer.twitter.com/en/docs/tweets/filter-realtime/guides/basic-stream-parameters#follow) for full details.
184
+
185
+
This command accepts one or more screen names.
186
+
187
+
You can feed it numeric Twitter user IDs instead of screen names by using the `--ids` flag.
188
+
189
+
The command also supports the `--sql` and `--attach` options, and the `--verbose` option for displaying tweets as they are captured.
190
+
191
+
Here's how to start following tweets from every user ID currently represented as being followed in the `following` table (populated using the `friends-ids` command):
192
+
193
+
$ twitter-to-sqlite follow tweets.db \
194
+
--sql="select distinct followed_id from following" \
195
+
--ids
196
+
149
197
## Design notes
150
198
151
199
* Tweet IDs are stored as integers, to afford sorting by ID in a sensible way
+1
-1
twitter_to_sqlite/cli.py
+1
-1
twitter_to_sqlite/cli.py
···
363
363
)
364
364
@click.option("--verbose", is_flag=True, help="Verbose mode: display every tweet")
365
365
def follow(db_path, identifiers, attach, sql, ids, auth, verbose):
366
-
"Experimental: Follow these Twitter users (numeric user IDs required) and save tweets in real-time"
366
+
"Experimental: Follow these Twitter users and save tweets in real-time"
367
367
auth = json.load(open(auth))
368
368
session = utils.session_for_auth(auth)
369
369
db = sqlite_utils.Database(db_path)