Consolidate file matches

Consolidate matches for a single file into a single output match file.

Use delimiters for each match:

```
#bed:begin {"dst":"/path/to/dst","pos":123","len":12}
This is my matched text
#bed:end
```

The issue names no files, tests, or entry points. Start by tracing how individual file matches are currently written, then identify where output can be grouped per source file. Done means one output match file contains all matches for each file, with the specified #bed:begin and #bed:end delimiters.

Written by the indexing model from the issue text.