liquidlight/tweetcam

Tweetcam - tweet a twitter account and, using a raspberry pi, get a photo or video back

tweetcam's Issues

Update last reply ID even if it's late

If the tweet is a delayed one, still update last reply ID

Update custom classes to be nicer and less clunky

Take inspiration from here:

https://github.com/seepel/bucketobytes

Add more things to config file

• How long between first tweet and recording start
• Media upload paths
• Whether to tweet before recording

Add more tweet text

Add more tweet text for photos and videos

Detect if gettingMentions throws error and handle nicely

TypeError: argument to reversed() must be a sequence

@mikestreety:
I have cloned your repo and installed all necessary imports (TwitterAPI, requests, etc.), I have changed _config.cfg to contain my own keys and did not touch any of the other default settings from the .skel file.

When I run python bot.py, as instructed by your README, I get the following error message from my Raspberry Pi:

Traceback (most recent call last):
  File "bot.py", line 10, in <module>
    for tweet in reversed(mentions):
TypeError: argument to reversed() must be a sequence

my bot.py looks like this:

#!/usr/bin/env python

# SETUP: Import global modules
import TweetStreetCam

g = TweetStreetCam.GraffCam()
mentions = g.GetMentions()

if mentions:
        for tweet in reversed(mentions):
                g.ActionTweet(tweet)

for tweet in g.GetStream():
        if 'text' in tweet:
                g.ActionTweet(tweet)

Im assuming this is a python problem right off the bat and might not have to do with my cloned version. The most apparent thing is that we are a TwitterResponse object back but the reversed method cannot handle it.

Do you know of a way to correct this?

Add Readme

Add instructions in a readme

Still graffcam references

https://github.com/liquidlight/tweetcam/blob/master/TweetStreetCam.py#L16

https://github.com/liquidlight/tweetcam/blob/master/bot.py#L6

Don't re-record failed uploads

Keep track of any failures and keep trying - rather than re-recording

Change hardcoded paths to dynamic

Python allows you to get the current path:

os.path.dirname(os.path.abspath(__file__))

Can replace this:

https://github.com/liquidlight/graffcam/blob/master/TweetStreetCam.py#L22

And there would then be no need for the path in the config file

Handle a 400 error - try and reupload

AVC-H264 import - frame size 1280 x 720 at 25.000 FPS
AVC Import results: 209 samples - Slices: 4 I 205 P 0 B - 0 SEI - 4 IDR
Saving /home/pi/tweet/media/videos/662227019011346433.mp4: 0.500 secs Interleaving
/usr/local/lib/python2.7/dist-packages/requests/packages/urllib3/util/ssl_.py:90: InsecurePlatformWarning: A true SSLContext object is not available. This prevents urllib3 from configuring SSL appropriately and may cause certain SSL connections to fail. For more information, see https://urllib3.readthedocs.org/en/latest/security.html#insecureplatformwarning.
  InsecurePlatformWarning
/usr/local/lib/python2.7/dist-packages/requests/packages/urllib3/util/ssl_.py:90: InsecurePlatformWarning: A true SSLContext object is not available. This prevents urllib3 from configuring SSL appropriately and may cause certain SSL connections to fail. For more information, see https://urllib3.readthedocs.org/en/latest/security.html#insecureplatformwarning.
  InsecurePlatformWarning
400
{"request":"\/1.1\/media\/upload.json","error":"segment size must be \u003C= 1."}

Add streaming

Rather than waiting up to a minute to send, it should use streaming to action and reply instantly

Delete successfully uploaded media

If a video or image is successfully uploaded, delete it from the file system

TypeError: string indices must be integers

Hi @mikestreety looks like my fix for #16 created another bug in its place; it is reproduced when the last_mention_id = 0. The resulting error is:

{u'errors': [{u'message': u'since_id parameter is invalid.', u'code': 44}]}
Traceback (most recent call last):
File "bot.py", line 7, in
mentions = g.GetMentions()
File "/home/amarzi/Documents/tweetcam/TweetStreetCam.py", line 64, in GetMentions
mentions = sorted(mentions, key=lambda k:k['id'])
File "/home/amarzi/Documents/tweetcam/TweetStreetCam.py", line 64, in
mentions = sorted(mentions, key=lambda k:k['id'])
TypeError: string indices must be integers

where the first line is the contents of the mentions response from twitter.

Can you try it on your system to see if you get the same bug/result?

I have a fix for it just in case; if twitter cannot somehow handle a 0 value for since_id then we can send in a parameterless mentions_timeline request like this: mentions = self.api.request('statuses/mentions_timeline', {}).json() if that value is 0, otherwise we will send a parametered mentions_timeline request like before: mentions = self.api.request('statuses/mentions_timeline', {'since_id': last_mention_id}).json()

The revised GetMentions() method would be as follows:

def GetMentions(self):
                last_mention_id = self.config.get('tweets', 'last_mention_id')
                if last_mention_id == str(0):
                        mentions = self.api.request('statuses/mentions_timeline', {}).json()
                else:
                        mentions = self.api.request('statuses/mentions_timeline', {'since_id': last_m$
                mentions = sorted(mentions, key=lambda k:k['id'])
                return mentions

Make getting tweet text a function of sorts

As it's all over the shop now:

start_status = random.choice(json.loads(self.config.get('tweet_text', 'preperation')))
                start_status = start_status.replace('[[user]]', '@%s' % (username))

Unify logging & correct typos

The logging is a bit all over the place and repeats itself (and misses the fact uploads fail...)

Restart process if process has died

Don't reply with media if tweet is old

If the tweet is more then X minutes old, rather than replying with a video/image, reply with an appology

Review readme

• Daemon file name is wrong (should be tweets) https://github.com/liquidlight/tweetcam#step-4-set-up-the-daemon
• Won't reply with a video if tweet difference limit is reached https://github.com/liquidlight/tweetcam#step-3-run-the-bot
• Copy, don't rename https://github.com/liquidlight/tweetcam#step-2-set-up-config-file

Create a config skel

For quick set up