Hi,

I'm rewriting my galaxy-api-tools project to use bioblend. I can't believe I missed it the first time around.

I'm a little confused reading about LibraryDatasetDatasetAssociation (ldda), LibraryDataset (ld) and HistoryDatasetAssociation (hda). I found something in the wiki[1] but is not helping enough.

I guess my question is, let's say I use any of bioblend.galaxy.libraries.upload_*() fucntions[2] to upload a file to a data library, the json response for that is something that looks like this: [{'url': '/api/libraries/ca4f6513e2849358/contents/e70b6e37a29da1c1', 'id': 'e70b6e37a29da1c1', 'name': '1.fastq'}]

Now let's say I use bioblend.galaxy.tools.upload_file()[2] and upload a file to a history, the json response for that looks like: {'outputs': [{'misc_blurb': None, 'peek': '<table cellspacing="0" cellpadding="3"></table>', 'uuid': None, 'hda_ldda': 'hda', 'data_type': 'auto', 'deleted': False, 'history_id': '6eafbc3efd020327', 'purged': False, 'state': 'queued', 'name': '1.fastq', 'visible': True, 'genome_build': '?', 'model_class': 'HistoryDatasetAssociation', 'hid': 5, 'file_size': 0, 'metadata_data_lines': None, 'file_ext': 'auto', 'id': 'c3e7863ed8b6cf9e', 'misc_info': None, 'metadata_dbkey': '?'}]}

Which 'src' setting should I use for bioblend.galaxy.workflows.run_workflow()'s[2] {'<input>': {'id': <encoded dataset ID>, 'src': '[ldda, ld, hda]'}?

If I have to guess {'<input>': {'id': 'e70b6e37a29da1c1', 'src': 'ldda'} for the first one and {'<input>': {'id': 'c3e7863ed8b6cf9e', 'src': 'hda'} for the second one. That leaves out 'ld', when do I have to use that one?

Finally and as a comment. I think the inconsistency on what to expect after delivering a payload to the API makes it hard to use these responses. See this issue[3] I submitted to bioblend but it seems is actually Galaxy who is at fault.

[1]http://wiki.galaxyproject.org/Admin/Config/Performance/Purge%20Histories%20a... [2]http://bioblend.readthedocs.org/en/latest/api_docs/galaxy/all.html [3]https://github.com/afgane/bioblend/issues/40

Thanks, Carlos