Raw label parser

[ejhumphrey]

tooling to convert CSV results from CrowdFlower into a sparse label format. Introduces num_responses into the table, and bounds relevance values on [0, 1]. Histogram confirms that there's no relevance scores in the middle, e.g. no relevance values in (0.25, 0.75):

[bmcfee]

Histogram confirms that there's no relevance scores in the middle, e.g. no relevance values in (0.25, 0.75):

Does it? I see some height down there. Can you plot on a log y axis to verify?

[ejhumphrey]

I checked the values directly, I'm confident there's nothing there

[coveralls]

Pull Request Test Coverage Report for Build 12

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 94.068%

---

Totals
Change from base Build 6:	0.0%
Covered Lines:	222
Relevant Lines:	236

---

💛 - Coveralls

[coveralls]

Pull Request Test Coverage Report for Build 12

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 94.068%

---

Totals
Change from base Build 6:	0.0%
Covered Lines:	222
Relevant Lines:	236

---

💛 - Coveralls