This change ensures that unknown tags with a defined VL are read as bytes (OW). This should fix #231. Previously they would have been read as strings by default.

Co-authored-by: Suyash Kumar <suyashkumar2003@gmail.com>

…#245)

This change introduces a ParseOption that allows PixelData element VL and expected lengths to mismatch without returning an error. If the given ParseOption is set, the PixelData with the aforementioned mismatch will have its raw bytes read, dropped into an encapsulated Frame, and have that Frame wrapped in a PixelDataInfo with a ParseErr property set to an error describing this scenario. 

(Commit description/summary set by suyashkumar@).

Co-authored-by: quang18 <quang@segmed.ai>

* chmod -x

* Documentation: http → https

This change wraps the read* methods in read.go into a lightweight reader struct. This allows for easier sharing of some common variables, for example the parse options set without having to inject them through deep read* call chains. 

This also does the following: 

* Moves readMetadata into read.go
*  removes parseOptSet from Dataset.
* updates dicomio.NewReader signature to not include an error (not needed). 


We may want to revisit the naming of some of these entities at some point. There now exists a dicom.reader (unexported) and a lower level dicomio.Reader (referred to as rawReader in this change). Ideally we can also make future refactors to eliminate the need for Parser to be aware of the rawReader (should be easy).

This introduces an option to SkipPixelData processing, making use of the new reader struct introduced in #254.

See also, review discussion in #243 that helped motivate the changes I wanted to make in #254.

…rving roundtrip read/write (#256)

This addresses _some_ of the discussion in #224. The SkipProcessingPixelDataValue option causes the PixelData bytes to be read into the element, but not processed any further. This option provides an option to save the CPU cycles processing NativePixel data but still gives a roundtrip-able Dataset.

If you want to save both CPU and Memory, you can instead use the SkipPixelData option, which doesn't load the PixelData bytes into memory at all.

In the future, we should consider an option to lazy load/process all Element values if possible.

Bumps [golang.org/x/text](https://github.com/golang/text) from 0.3.7 to 0.3.8.
- [Release notes](https://github.com/golang/text/releases)
- [Commits](golang/text@v0.3.7...v0.3.8)

…s. (#261)

This change updates the Parse benchmark to operate across useful ParseOptions, adds memory benchmarking, and updates some benchmark params. 

For example, we can take a look at the CPU / memory differences across the new ParseOptions for one of the test dicoms:
```
BenchmarkParse/NoOptions/4.dcm-4                                     	      10	 106876945 ns/op	120572543 B/op	    2160 allocs/op
BenchmarkParse/SkipPixelData/4.dcm-4                                 	      10	    531193 ns/op	   34900 B/op	    2144 allocs/op
BenchmarkParse/SkipProcessingPixelDataValue/4.dcm-4                  	      10	   1807256 ns/op	 7573518 B/op	    2146 allocs/op
```