Add CURIE parse benchmark (#499)

* Add CURIE parse benchmark

* Update pages

* Add more epochs to benchmark

* Update README.md

* Pass mypy

exports/benchmarks/README.md

@@ -2,3 +2,12 @@
 
 1. The [`uri_parsing`](uri_parsing) benchmark checks the `bioregistry.parse_iri`
    function. See also https://github.com/biopragmatics/bioregistry/pull/481.
+2. The [`curie_parsing`](curie_parsing) benchmark checks
+   the `bioregistry.parse_curie`
+   function.
+
+## Overview
+
+| URI Parsing                  | CURIE Parsing                  |
+|------------------------------|--------------------------------|
+| ![](uri_parsing/results.svg) | ![](curie_parsing/results.svg) |

exports/benchmarks/curie_parsing/README.md

@@ -0,0 +1,32 @@
+# CURIE Parsing Benchmark
+
+This benchmark checks the `bioregistry.parse_curie` function.
+
+## Dataset
+
+The benchmarking dataset is available in [`data.tsv`](data.tsv). It contains
+5 columns:
+
+1. `prefix` - a canonical Bioregistry prefix
+2. `identifier` - a local unique identifier in the prefix's semantic space
+3. `prefix_synonym` - the synonym of the canonical prefix being used
+4. `banana` - the banana being used (i.e., redundant prefix in local identifier)
+5. `curie` - the CURIE for the prefix_synonym-banana-identifier triple
+
+Example data:
+
+| prefix        | identifier   | prefix_synonym | banana | curie                      |
+|---------------|--------------|----------------|--------|----------------------------|
+| 3dmet         | B00162       | 3dmet          |        | 3dmet:B00162               |
+| 4dn.biosource | 4DNSR73BT2A2 | 4DN            |        | 4DN:4DNSR73BT2A2           |
+| 4dn.biosource | 4DNSR73BT2A2 | 4dn.biosource  |        | 4dn.biosource:4DNSR73BT2A2 |
+| 4dn.replicate | 4DNESWX1J3QU | 4dn.replicate  |        | 4dn.replicate:4DNESWX1J3QU |
+| abcd          | AD834        | abcd           |        | abcd:AD834                 |
+| cco           | 0000003      | cco            |        | cco:0000003                |
+| cco           | 0000003      | cco            | CCO:   | cco:CCO:0000003            |
+
+## Results
+
+Most parsing goes pretty fast (average around 4,000 CURIE/second).
+
+![](results.svg)