The relevant section:
You may, however, provide export via non-automated means (e.g., download of spreadsheets or PDF files, or use of a “save as” button) of up to 50,000 public Tweets and/or User Objects per user of your Service, per day.
We’ve got a small dataset (approximately 3500 tweets) that has been human annotated that we’d like to redistribute (we’re working on cross-domain sentiment analysis). Unfortunately, the original tweet identifiers are mangled, and so our only option is to instead try and manually search for each tweet and match up the identifiers. Would 6) b) i) allow us to redistribute the raw tweet text alongside the annotations, provided we limited the total number of downloads?