rra/dafr
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

crawler now looks for instance metadata, started to abstract collection into functions

Browse Source
This commit is contained in:
rra 2018-05-30 10:27:02 +02:00
parent d90366d4c6
commit 4cac59e445
1 changed files with 29 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

42
fedicrawler.py
View File

@ -1,30 +1,45 @@
#!/bin/env python3 #!/bin/env python3
# fediscraper v1 # fediscraper v1
import json, requests import json, requests,threading
start_url = 'https://post.lurk.org' start_url = 'https://post.lurk.org'
activity = '' activity = ''
peers_info = '/api/v1/instance/peers' peers_info = '/api/v1/instance/peers'
instance_info = '/api/v1/instance' instance_info = '/api/v1/instance'
instances = {set([])} def get_peers(instance):
r = requests.get('https://'+instance+'/api/v1/instance/peers', timeout=10)
if r.status_code == 200:
peers = r.json()
print(instance, 'peers with', len(peers))
return(peers)
else:
print('fail: got {} on {}'.format(r.status_code, instance))
instances = {}
r = requests.get(start_url+peers_info) r = requests.get(start_url+peers_info)
if r. status_code == 200: if r. status_code == 200:
print('200 for '+start_url) print('200 for '+start_url)
peers = r.json() peers = r.json()
print('{} has {} peers'.format(start_url, len(peers))) print('{} has {} peers'.format(start_url, len(peers)))
for count, peer in enumerate(peers): for count, peer in enumerate(peers):
instances.add(peer) #instances.add(peer)
try: try:
r = requests.get('https://'+peer+peers_info, timeout=10) peer_list = get_peers(peer)
if r.status_code == 200: if peer_list:
print(peer, 'peers with', len(r.json())) for i in peer_list:
for i in r.json():
if i not in instances: if i not in instances:
instances.add(i) try:
print('added {}, n={}'.format(i,len(instances))) ii = requests.get('https://'+i+instance_info, timeout=10)
info = ii.json()
except Exception as e:
print('failed to query instance info')
print(e)
info = 'error'
pass
instances[i] = info
print('added {}, n={}'.format(i,len(instances)))
else: else:
print(i.status_code, 'on', peer) print(i.status_code, 'on', peer)
except Exception as e: except Exception as e:
@ -32,7 +47,8 @@ if r. status_code == 200:
# instances[peer] = {'error':e} # instances[peer] = {'error':e}
print(e) print(e)
text = list(filter(None.__ne__, instances)) #text = list(filter(None.__ne__, instances))
with open('instance_scrape.txt','w') as f: with open('instance_scrape.json','w') as f:
f.write('\n'.join(text)) json.dumps(instances,f)
#f.write('\n'.join(text)