Create hocr parser

[dcloud]

Found existing one in Github, but it didn't work. See if we can quick write one of our own.

[dcloud]

Looking at https://gist.github.com/dcloud/9173113, fwiw

[dcloud]

Basics done in 9036cdc.

[jsfenfen]

Note that spans for words are sometimes ocrx_word and sometimes just ocr_word -- in other words, the x is sometimes missing.

[dcloud]

Ah, I wasn't sure about that. Reopening. Do you know the difference (what the x means)?

[jsfenfen]

I dunno. I'm not sure it's intentional or a bug. But since I ran into this
I've gotten more skeptical about how tight the spec is...

On Mon, Feb 24, 2014 at 12:22 PM, Daniel Cloud [email protected]:

Ah, I wasn't sure about that. Reopening. Do you know the difference (what
the x means)?

Reply to this email directly or view it on GitHubhttps://github.com//issues/1#issuecomment-35910887
.

[jsfenfen]

Yeah, so fwiw ocrx_word might not be a formal part of the spec -- this doc
https://docs.google.com/document/d/1QQnIQtvdAC_8n92-LhwPcjtAUFwBlzE8EWnKAxlgVf0--
I'm not totally sure of how authoritative it is -- describes it as
being
part of the 'engine-specific markup'. Which gives me pause...

On Mon, Feb 24, 2014 at 12:22 PM, Daniel Cloud [email protected]:

Reopened #1 #1.

Reply to this email directly or view it on GitHubhttps://github.com//issues/1
.