bellingcat/snscrape
1
0
Fork 0
mirror of https://github.com/bellingcat/snscrape.git synced 2026-07-04 14:28:36 +03:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #218 from NoeCampos22/Place_Data

Browse Source 
Extract more information on Twitter places
This commit is contained in:
JustAnotherArchivist
2021-04-20 01:45:22 +00:00
committed by GitHub
parent 39a34a57ac 8f1c470061
commit a34f93076a
1 changed files with 11 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
snscrape/modules/twitter.py
View File

@@ -42,7 +42,7 @@ class Tweet(snscrape.base.Item):
quotedTweet: typing.Optional['Tweet'] = None quotedTweet: typing.Optional['Tweet'] = None
mentionedUsers: typing.Optional[typing.List['User']] = None mentionedUsers: typing.Optional[typing.List['User']] = None
coordinates: typing.Optional['Coordinates'] = None coordinates: typing.Optional['Coordinates'] = None
place: typing.Optional[str] = None place: typing.Optional['Place'] = None
username = snscrape.base._DeprecatedProperty('username', lambda self: self.user.username, 'user.username') username = snscrape.base._DeprecatedProperty('username', lambda self: self.user.username, 'user.username')
outlinksss = snscrape.base._DeprecatedProperty('outlinksss', lambda self: ' '.join(self.outlinks), 'outlinks') outlinksss = snscrape.base._DeprecatedProperty('outlinksss', lambda self: ' '.join(self.outlinks), 'outlinks')
@@ -99,6 +99,15 @@ class Coordinates:
latitude: float latitude: float
@dataclasses.dataclass
class Place:
fullName: str
name: str
type: str
country: str
countryCode: str
@dataclasses.dataclass @dataclasses.dataclass
class User(snscrape.base.Entity): class User(snscrape.base.Entity):
# Most fields can be None if they're not known. # Most fields can be None if they're not known.
@@ -378,7 +387,7 @@ class TwitterAPIScraper(snscrape.base.Scraper):
if (coords := tweet['geo']['coordinates']) and len(coords) == 2: if (coords := tweet['geo']['coordinates']) and len(coords) == 2:
kwargs['coordinates'] = Coordinates(coords[1], coords[0]) kwargs['coordinates'] = Coordinates(coords[1], coords[0])
if tweet.get('place'): if tweet.get('place'):
kwargs['place'] = tweet['place']['full_name'] kwargs['place'] = Place(tweet['place']['full_name'], tweet['place']['name'], tweet['place']['place_type'], tweet['place']['country'], tweet['place']['country_code'])
if 'coordinates' not in kwargs and tweet['place']['bounding_box'] and (coords := tweet['place']['bounding_box']['coordinates']) and coords[0] and len(coords[0][0]) == 2: if 'coordinates' not in kwargs and tweet['place']['bounding_box'] and (coords := tweet['place']['bounding_box']['coordinates']) and coords[0] and len(coords[0][0]) == 2:
# Take the first (longitude, latitude) couple of the "place square" # Take the first (longitude, latitude) couple of the "place square"
kwargs['coordinates'] = Coordinates(coords[0][0][0], coords[0][0][1]) kwargs['coordinates'] = Coordinates(coords[0][0][0], coords[0][0][1])