Feature request: Limit the number of documents to parse and classify

[ggozad]

Feature request: Limit the number of documents to parse and classify

[ggozad]

While this is doable and I would like it to be used, it needs some brushing up:
What are the criteria for selection? An absolute number will pick random documents, and may not parse enough to be accurate...
Maybe an async queue would be a better idea?

Also I would like if possible to have some benchmarks. How many documents are we talking about? How long does it take to parse them all? Did you perform the test with the n-gram tagger? The Pen treebank is too slow anyway...