KaedeSkirata wrote:Hey!
Few issues so far - though the bot is badass.
I want to set up a "watch for keyword" stream and do different things for each tagged keyword
Just as a warning, it is quite a complex module and has not been as widely used as I expected - although I think it has huge potential (especially for email scraping, easily grab 1000s a day of highly targeted emails. Anyway, due to the lack of users using, there is probably still some issues (as you appear to have found) and lots of room for feedback, so please do report / suggest any and everything
so onto your issue
Code: Select all
Unable to cast object of Type 'LibTwitter.Logic.Filter.FilterUser' to type 'LibTwitter.Logic.Filter.FilterUserWithSiteScrape
Can you submit logs, this should be a simple fix with the logs
. E.G.
for #kings I want to like and follow everyone,
for #queens I want to like everything and
for #joker (VERY high volume) I want to like/follow one person per minute (that's heavily filtered down).
I'd also like to scrape all the Twitter ID's for everyone in any of the hashtags, which leads me to my issues.
ok this should be doable, the scrape actions (ids / urls / details / emails happens on all responses, whether the action (like / follow..) are less often
As stated at end of this reply, I was mistaken there is an ignore actions on keyword function so you don't end up trying to follow users on a high traffic keyword a 1000 in a minute, I was wrong there isn't
Issue #1 - In the "watch for keyword" stream, everytime I try and add a user filter (language, profile pic, etc.) I get an error from the bot saying it was
Code: Select all
Unable to cast object of Type 'LibTwitter.Logic.Filter.FilterUser' to type 'LibTwitter.Logic.Filter.FilterUserWithSiteScrape
ah ok, this gives me a bit more detail to go on, let me check around this area to see if I can
a) replicate
b) fix
without the need for your logs.
Ok, have found the issue - am fixing now, update will be ready I hope later today
Issue #2 - (I think I might have fixed this, but putting it up anyways) Occasionally, when restarting the stream, under the details of the process on the processing tab it would be spamming "Error:Stream Paused/Stream Restarted"
and your "fix"? This is just logging the connection to the twitter stream server. if it is thrasing pause / restart / pause restart, then something is not quite right. More details (as always) should be in the logs
HELP > LOGS > SUBMIT
then send your logs ID - the first 4 numbers is sufficient (displayed after successful uploading of logs)
Issue #3 - When I was running a stream JUST to scrape details, I kept getting an "Error: Value cannot be null" (paraphrasing from memory)
Again, logs would help. But it is basically the programming trying to do something with an expected value that doesn't exist.
Issue #4 - I have 2 accounts I'm testing right now - in totally separate niches. For one, I had a stream of related hashtags setup (e.g. #kings, #queens, etc.) but when I went to add a NEW stream for the second account, there was no option to get rid of the previous stream selections. All I could do was uncheck the 'action' section and let it run.
ah ok, like a clear all / start anew option you want?
Which brings me to my questions.
Q1 - Does scraping twitter user details from the stream count against my API limit? Does each individual filter call do the same?
No, that is the beauty of the function. Set up a stream of 2000 kws (I think that is the limit) and scrape all details to your hearts content, without effecting your API limits at all
Q2 - Does scraping a twitter user's details from the stream count as 'handling' a tweet for the purposes of the pause timer?
scraping details is outside of any timings. It is only actions (like / follow / retweet / etc ) that are effected (as scraping is local, the data has been received by the stream so might as well grab it all. Actions such as liking may have to be limited due to traffic of keyword
Q3 - Is there a way to force a pause only for certain hashtags in the stream? Ideally, I want a uniform pause between follow attempts on high-volume keywords, but for the smaller ones (brandname, for instance) I want to ensure that they are all engaged with immediately
I was sure there was, but checking there isn't, I think I am confusing this with the WATCH FOR NEW FOLLOWERS STREAM, where you can action once every X minutes. I will have to add the ability to ignore tweets that come in on a per keyword basis. I am currently frantically updating the instadub (ISG) but due to change by ISG breaking lots of functionality. If can remind me about this tomorrow / Friday
Sorry for the great wall of text!
