New read length definition, improved docs and error handling

Minor update:

• Read length is now defined as the number of cycles run for a read. Read length intervals in config have been updated as an effect of this.
• Read length intervals are now lower and upper inclusive. This together with the new read length defintion solved issue #53 .
• Config usage have been moved from config to docs.
• The issue where a certain machine type, run type and read length combination can't be found in the config, will now be communicated to the user.

Improved documentation, JOSS paper and bug fixes

This release contains:

• Our paper submitted to JOSS on CheckQC
• Many improvements to the documentation, including building the docs using ReadTheDocs. The docs are now available at http://checkqc.readthedocs.io/en/latest/
• Simplified installation process, where Interop does not need to be installed manually prior to CheckQC
• The reads per sample handler did not report the threshold that was actually used as reported in #49. This has been fixed.
• The location of the bcl2fastq output directory is now configurable, with the default being the default used by bcl2fastq. #47

More information in log output, and better exception handling

This minor release updates a few things:

• The log output will now include information on what type of runfolder was identified, which handlers are activated and how they are configured.
• It has been clarified in the README that we required python 3.5+. If there are a lot of people requesting support for older python version that is something that we might have to look into, but for now this is how things are.
• Better exception handling behind the scenes which should avoid some ugly stack traces, and make the error messages more user friendly (e.g. reminding the user that they have to run bcl2fastq before running checkQC if there is no Stats.json file).

These updates were based on the feedback from @avilella, @af8, @senthil10 and @ewels. Thanks!

Working version flag and includes default the logger config in distribution

• Fixes a bug in the way that the version flag was handled. It is now possible to use it without also specifying a runfolder.
• Improves the install process by excluding the tests, and including the default logging config in the distribution.

Monitored path for webservice must exist and dynamic importing of handlers working

This release fixes two bugs:

• The monitored path for the webservice is now validated as the service is started
• A bug in the dynamic importing of the handlers meant the it was not possible to start CheckQC from any directory. This has now been fixed.

Webservice, json output mode and more!

This release adds two major features. The first is a new way of running CheckQC - now it is possible to run it as a webservice as well as a normal command line application.

To run CheckQC as a webservice, use the following command:

checkqc-ws <path to directory with runfolders>

The second feature is a json-mode for the cli. Adding --json when running CheckQC will print the output of the program as json to stdout. This is e.g. useful if you want to pass the CheckQC output to other programs to futher parse the data.

In addition to these feature the following minor features are also part of this relase.

• Validation of handler config
• CheckQC will now raise an error if there is no error rate present in a run (e.g. because no PhiX was added to the lane). This behaviour can be overridden by setting allow_missing_error_rate: True in the ErrorRateHandler part of the config.
• Added commandline option to print the program version

Enjoy!

First release of CheckQC

This is the 1.0.0 release of CheckQC. 🎆

Running CheckQC is as easy as:

checkqc <path to your runfolder>

In this version CheckQC supports the following Illumina instruments:

• HiSeqX
• HiSeq2500
• MiSeq
• NovaSeq

It has five handlers which allows examining the following QC metrics:

• UndeterminedPercentageHandler, checks the amount of reads which could not be assigned to a specific sample.
• ClusterPFHandler, checks the amount of clusters (reads) which have passed Illuminas filters.
• Q30Handler, checks the percentage of data which is above Q30.
• ErrorRateHandler, checks that the error rate is below a set value
• ReadsPerSampleHandler, checks that the number of reads assigned to a sample is not below what one would expect given the number of samples on that lane (assuming that all samples pooled together should have the same amount of data)

Enjoy!

v1.0.0-rc1

Merge pull request #26 from johandahlberg/validation_testing

Fixing minor things found when preparing for validation testing