README.md

# Validation of Metadata Documents

## Purpose

After you generated a new *Process* file, it is desirable to check that the document 
adheres to the DEEP *Process* specification that is defined in the XML schema file `template/deep_process_schema.xsd`.
To achieve that with little to no extra effort, you can run the small Python validation script `mdvalid.py`.

## Installation
The script itself requires no installation or configuration and works with Python2.7+ or Python3.2+.
It uses the *lxml* package that is not part of the Python standard library.
Please install [lxml](http://lxml.de/installation.html) according to your software environment.
For your convenience, we provide two YAML files to install a complete
environment for Python2 or Python3 (whichever you prefer)
using the [Conda](http://conda.pydata.org/miniconda.html) package manager.

---

#### Attention

Note from: 2016-12-29

Due to some dependency issues concerning the *lxml* package and Conda, *lxml* currently it has to be installed
from within the activated Conda environment. That is why, in the Conda YAML files,
the line `- lxml=3.*` is commented out - see also here: [Github issue 4093](https://github.com/conda/conda/issues/4093)

---


In case you want to use the Conda package manager, you can create the environment as follows:

```bash
$ conda env create -f conda_py3_mdvalid.yml
## wait for setup to complete
$ source activate py3valid
## next line: workaround solution - 2016-12-29
$ pip install lxml
## wait for installation to complete
## run the script; see below
$ source deactivate
```

## Execution

The script can be executed on a command line as follows:

```bash
$ python mdvalid.py [Options]
```

Please read the help for details:

```bash
$ python mdvalid.py --help
```

The Python code has been developed on a Debian Linux system:

* x86_64 Linux (Debian 7.5 "Wheezy")

with Python versions

* Python 2.7.3 (lxml.etree 2.3.2)
* Python 3.2.3 (lxml.etree 2.3.2)

Note that this testing setup refers to the released version. The current version of *lxml* 3.6.4 works as well.

## Examples:

Validate a process XML:
```bash
$ python mdvalid.py --schema schema_file.xsd --process process_file.xml
```

Validate a process XML and a corresponding analysis metadata file:
```bash
$ python mdvalid.py --schema schema_file.xsd --process process_file.xml --analysis analysis_file.amd.tsv
```

Validate many analyses of the same type:
```bash
$ python mdvalid.py --schema schema_file.xsd --process process_file.xml --analysis file1.amd.tsv file2.amd.tsv file3.amd.tsv
```

In case of errors, you can get more info by running the script in debug mode
```bash
$ python mdvalid.py --debug [Options]
```