balkian/bitter
1
0
Fork 0
mirror of https://github.com/balkian/bitter.git synced 2024-12-21 08:08:12 +00:00
Code Issues Releases Wiki Activity

Version 0.10.3

Browse Source 
* Ability to include "optional" fields from tweets (e.g., retweeted_status).
* Optional caching (for very large datasets)
This commit is contained in:
J. Fernando Sánchez 2020-06-24 11:01:02 +02:00
parent 030c41b826
commit ea848f1a78
4 changed files with 93 additions and 60 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
bitter/VERSION
View File

@ -1 +1 @@
0.10.2
0.10.3

6
bitter/__init__.py
View File

@ -6,10 +6,12 @@ http://github.com/balkian/bitter
import os
from .version import __version__
from . import config as bconf
def easy(*args, **kwargs):
def easy(conffile=bconf.CONFIG_FILE):
from .crawlers import TwitterQueue
return TwitterQueue.from_credentials(*args, **kwargs)
return TwitterQueue.from_config(conffile=conffile)
__all__ = ['cli', 'config', 'crawlers', 'models', 'utils' ]

74
bitter/cli.py
View File

@ -8,8 +8,6 @@ import time
import sqlalchemy.types
import threading
import sqlite3
import operator
from functools import reduce
from tqdm import tqdm
from sqlalchemy import exists
@ -19,7 +17,6 @@ from bitter import config as bconf
from bitter.models import make_session, User, ExtractorEntry, Following
import sys
import csv as tsv
if sys.version_info <= (3, 0):
from contextlib2 import ExitStack
else:
@ -29,48 +26,29 @@ else:
logger = logging.getLogger(__name__)
def serialize(function):
'''Common options to serialize output to CSV or other formats'''
@click.option('--csv', help='Print each object as a csv row. Provide a list of comma-separated fields to print.', default='', type=str)
@click.option('--fields', help='Provide a list of comma-separated fields to print.', default='', type=str)
@click.option('--ignore_missing', help='Do not show warnings for missing fields.', is_flag=True)
@click.option('--header', help='Header that will be printed at the beginning of the file', default=None)
@click.option('--csv', help='Print each object as a csv row.', is_flag=True)
@click.option('--jsonlines', '--json', help='Print each object as JSON in a new line.', is_flag=True)
@click.option('--indented', help='Print each object as an indented JSON object', is_flag=True)
@click.option('--outdelimiter', help='Delimiter for some output formats, such as CSV. It defaults to \t', default='\t')
@click.option('--outfile', help='Output file. It defaults to STDOUT', default=sys.stdout)
def decorated(csv, header, jsonlines, indented, outfile, **kwargs):
if header:
print(header)
def decorated(fields, ignore_missing, header, csv, jsonlines, indented, outfile, outdelimiter, **kwargs):
it = function(**kwargs)
outformat = 'json'
if csv:
outformat = 'csv'
elif jsonlines:
outformat = 'jsonlines'
elif indented:
outformat = 'indented'
def do(out):
return utils.serialized(it, outfile, outformat=outformat, fields=fields.split(','), ignore_missing=ignore_missing, header=header, delimiter=outdelimiter)
if csv:
delimiter = '\t'
writer = tsv.writer(out, quoting=tsv.QUOTE_ALL, delimiter=delimiter)
if header is None:
# Print fields as header unless told otherwise
print(csv.replace(',', delimiter), file=out)
fields = list(token.strip().split('.') for token in csv.split(','))
for obj in it:
writer.writerow(list(reduce(operator.getitem, field, obj) for field in fields))
elif jsonlines:
for obj in it:
print(json.dumps(obj, sort_keys=True), file=out)
elif indented:
for obj in it:
print(json.dumps(obj, indent=4, sort_keys=True), file=out)
else:
for obj in it:
print(obj, file=out)
if outfile is sys.stdout:
return do(sys.stdout)
with open(outfile, 'w') as out:
return do(out)
return decorated
@ -190,13 +168,14 @@ The result is stored as individual json files in your folder of choice.''')
@click.option('-u', '--update', is_flag=True, default=False, help='Download tweet even if it is already present. WARNING: it will overwrite existing files!')
@click.option('-r', '--retry', is_flag=True, default=False, help='Retry failed downloads')
@click.option('-d', '--delimiter', default=",")
@click.option('-n', '--nocache', is_flag=True, default=False, help='Do not cache results')
@click.option('--skip', help='Discard the first DISCARD lines (use them as a header)', default=0)
@click.option('--commentchar', help='Lines starting with this character will be ignored', default=None)
@click.option('-q', '--quotechar', default='"')
@click.option('-c', '--column', type=int, default=0)
@serialize
@click.pass_context
def get_tweets(ctx, tweetsfile, folder, update, retry, delimiter, skip, quotechar, commentchar, column):
def get_tweets(ctx, tweetsfile, folder, update, retry, delimiter, nocache, skip, quotechar, commentchar, column):
if update and not click.confirm('This may overwrite existing tweets. Continue?'):
click.echo('Cancelling')
return
@ -204,10 +183,9 @@ def get_tweets(ctx, tweetsfile, folder, update, retry, delimiter, skip, quotecha
status = tqdm('Queried')
failed = 0
for tid, obj in utils.download_file(wq, tweetsfile, folder, delimiter=delimiter,
batch_method=utils.tweet_download_batch,
skip=skip, quotechar=quotechar, commentchar=commentchar,
column=column, update=update, retry_failed=retry):
for tid, obj in utils.download_tweets_file(wq, tweetsfile, folder, delimiter=delimiter, cache=not nocache,
skip=skip, quotechar=quotechar, commentchar=commentchar,
column=column, update=update, retry_failed=retry):
status.update(1)
if not obj:
failed += 1
@ -264,6 +242,7 @@ def get_user(user, dry_run, folder, update):
@click.option('-f', '--folder', default="users")
@click.option('-u', '--update', is_flag=True, default=False, help='Download user even if it is already present. WARNING: it will overwrite existing files!')
@click.option('-r', '--retry', is_flag=True, default=False, help='Retry failed downloads')
@click.option('-n', '--nocache', is_flag=True, default=False, help='Do not cache results')
@click.option('-d', '--delimiter', default=",")
@click.option('--skip', help='Discard the first SKIP lines (e.g., use them as a header)',
is_flag=True, default=False)
@ -272,17 +251,17 @@ def get_user(user, dry_run, folder, update):
@click.option('-c', '--column', type=int, default=0)
@serialize
@click.pass_context
def get_users(ctx, usersfile, folder, update, retry, delimiter, skip, quotechar, commentchar, column):
def get_users(ctx, usersfile, folder, update, retry, nocache, delimiter, skip, quotechar, commentchar, column):
if update and not click.confirm('This may overwrite existing users. Continue?'):
click.echo('Cancelling')
return
wq = crawlers.TwitterQueue.from_config(conffile=bconf.CONFIG_FILE)
for i in utils.download_file(wq, usersfile, folder, delimiter=delimiter,
batch_method=utils.user_download_batch,
update=update, retry_failed=retry,
skip=skip, quotechar=quotechar,
commentchar=commentchar,
column=column):
for i in utils.download_users_file(wq, usersfile, folder, delimiter=delimiter,
update=update, retry_failed=retry,
skip=skip, quotechar=quotechar,
cache=not nocache,
commentchar=commentchar,
column=column):
yield i
@users.command('crawl')
@ -480,7 +459,6 @@ def extract(ctx, recursive, user, name, initfile):
@extractor.command('reset')
@click.pass_context
def reset_extractor(ctx):
wq = crawlers.TwitterQueue.from_config(conffile=bconf.CONFIG_FILE)
db = ctx.obj['DBURI']
session = make_session(db)
session.query(ExtractorEntry).filter(ExtractorEntry.pending==True).update({'pending':False})

71
bitter/utils.py
View File

@ -19,7 +19,9 @@ import queue
import threading
from select import select
from functools import partial
import operator
from functools import partial, reduce
from tqdm import tqdm
@ -473,22 +475,22 @@ def get_user(c, user):
except ValueError:
return c.users.lookup(screen_name=user)[0]
def download_tweet(wq, tweetid, write=True, folder="downloaded_tweets", update=False):
def download_tweet(wq, tweetid, cache=True, folder="downloaded_tweets", update=False):
tweet = cached_id(tweetid, folder)
if update or not tweet:
tweet = get_tweet(wq, tweetid)
if write and update:
if cache and update:
if tweet:
js = json.dumps(tweet)
write_json(js, folder)
yield tweet
def download_user(wq, userid, write=True, folder="downloaded_users", update=False):
def download_user(wq, userid, cache=True, folder="downloaded_users", update=False):
user = cached_id(userid, folder)
if update or not user:
user = get_user(wq, userid)
if write and update:
if cache and update:
if user:
write_json(user, folder, aliases=[user['screen_name'], ])
yield user
@ -589,7 +591,8 @@ def dump_result(oid, obj, folder, ignore_fails=True):
with open(fail_file(oid, folder), 'w') as f:
print('Object not found', file=f)
def download_list(wq, lst, folder, update=False, retry_failed=False, ignore_fails=False,
def download_list(wq, lst, folder, update=False, retry_failed=False, ignore_fails=False, cache=True,
batch_method=tweet_download_batch):
done = Queue()
@ -647,13 +650,24 @@ def download_list(wq, lst, folder, update=False, retry_failed=False, ignore_fail
break
oid, obj = rec
dump_result(oid, obj, folder, ignore_fails)
if cache or (not obj):
dump_result(oid, obj, folder, ignore_fails)
yield rec
wait.join()
def download_file(wq, csvfile, folder, column=0, delimiter=',', skip=0,
def download_tweets_file(*args, **kwargs):
kwargs['batch_method'] = tweet_download_batch
yield from download_file(*args, **kwargs)
def download_users_file(*args, **kwargs):
kwargs['batch_method'] = user_download_batch
yield from download_file(*args, **kwargs)
def download_file(wq, csvfile, folder, column=0, delimiter=',', skip=0, cache=True,
quotechar='"', commentchar=None, batch_method=tweet_download_batch,
**kwargs):
with open(csvfile) as f:
@ -670,7 +684,7 @@ def download_file(wq, csvfile, folder, column=0, delimiter=',', skip=0,
yield row[column].strip()
for res in download_list(wq, reader(csvreader), folder, batch_method=batch_method,
for res in download_list(wq, reader(csvreader), folder, batch_method=batch_method, cache=cache,
**kwargs):
yield res
@ -748,3 +762,42 @@ def _users_control(func, apiargs, remaining=0, **kwargs):
if int(cursor) != -1:
stop = False
return resp['users'], stop
def serialized(it, outfile, outformat='csv', fields=[], header=None, ignore_missing=False, delimiter='\t'):
outformat = outformat.lower()
def do(out):
if outformat == 'csv':
writer = csv.writer(out, quoting=csv.QUOTE_ALL, delimiter=delimiter)
if header != '':
h = header
if h is None:
h = delimiter.join(fields)
print(h, file=out)
attrs = list(token.strip().split('.') for token in fields)
for obj in it:
values = []
for attr in attrs:
try:
values.append(reduce(operator.getitem, attr, obj))
except KeyError:
if not ignore_missing:
print('Key not present: {}'.format(attr), file=sys.stderr)
values.append(None)
writer.writerow(values)
elif outformat == 'jsonlines':
for obj in it:
print(json.dumps(obj, sort_keys=True), file=out)
elif outformat == 'indented':
for obj in it:
print(json.dumps(obj, indent=4, sort_keys=True), file=out)
else:
for obj in it:
print(obj, file=out)
if outfile is sys.stdout:
return do(sys.stdout)
with open(outfile, 'w') as out:
return do(out)