Add MSmetaEnhancer #199

xtrojak · 2021-12-02T15:51:59Z

This PR adds MSmetaEnhancer. It is possible to select which conversion jobs will executed. The list of jobs is in a macros files generated dynamically from current version of MSMetaEnhancer (#197). It has to be updated upon updating this wrapper as described in readme file. The jobs can be specified in particular order (the performance of tool depends on that, #198) and/or in a bulk of jobs to be executed once the ordered ones are finished.

Close #176.
Close #198.
Close #197.
Close #200.

tools/msmetaenhancer/msmetaenhancer.xml

tools/msmetaenhancer/test-data/sample.msp

xtrojak · 2021-12-06T12:18:37Z

After discussions with @martenson and @hechth, we decided to proceed as following:

regarding MSMetaEnhancer: list of options as input #198, there does not seem to be an optimal solution. Therefore we decided to combine the two suggested solutions. The user can specify order of jobs using single select in repeat tag and/or unordered jobs in multiple select tag.
regarding MSMetaEnhancer: Dynamic conversion options #197, I have enumerated the options in macro file, which seems to be the easiest approach.

hechth · 2021-12-09T09:03:11Z

I'm currently trying the tool and it seems like it doesn't actually start any jobs and just terminates straight away.

martenson · 2021-12-09T09:47:18Z

@hechth any errors in the logs? The job finishes for me.

martenson

I think the tool would benefit from some more clarity in its interface. E.g. the most basic usage option probably is "do not select any jobs, just let MSME run everything" -- but that option is more or less hidden atm even though it is mentioned in the end of the help text.

I am also confused why the results are different for "no conversion jobs selected" and "all conversion jobs selected as a part of the random selector". I'd expect them be the same but they are consistently different.

For UX I'd consider a two-modes-tool with "simple" having no parameters other than the input file, and "specify jobs" offering the interface to alter the order/jobs.

tools/msmetaenhancer/msmetaenhancer.xml

hechth · 2021-12-09T10:30:50Z

The job finishes but just doesn't do anything and the input file doesn't change.

hechth · 2021-12-09T13:38:37Z

I think the tool would benefit from some more clarity in its interface. E.g. the most basic usage option probably is "do not select any jobs, just let MSME run everything" -- but that option is more or less hidden atm even though it is mentioned in the end of the help text.

I am also confused why the results are different for "no conversion jobs selected" and "all conversion jobs selected as a part of the random selector". I'd expect them be the same but they are consistently different.

For UX I'd consider a two-modes-tool with "simple" having no parameters other than the input file, and "specify jobs" offering the interface to alter the order/jobs.

For me the UI is completely fine and understandable and makes sense as it is to be honest.

xtrojak · 2021-12-09T15:10:36Z

The job finishes but just doesn't do anything and the input file doesn't change.

There was really an issue in order of arguments in job_options macro. Should be fixed in 24ae942.

xtrojak · 2021-12-09T15:30:00Z

I think the tool would benefit from some more clarity in its interface. E.g. the most basic usage option probably is "do not select any jobs, just let MSME run everything" -- but that option is more or less hidden atm even though it is mentioned in the end of the help text.

I suppose we can improve this in the future if the tool will not be intuitive enough to the users.

I am also confused why the results are different for "no conversion jobs selected" and "all conversion jobs selected as a part of the random selector". I'd expect them be the same but they are consistently different.

This should be covered in version of MSME 0.1.1 (changed in da914e9). At least I'm not able to reproduce it. Please provide sample .msp file if this issue is still lingering.

hechth · 2021-12-10T12:23:54Z

I found another bug when running only a single job (Pubchem: name -> inchikey)

Traceback (most recent call last):
  File "/m2b/home/hechth/git/recetox/galaxytools/tools/msmetaenhancer/msmetaenhancer_wrapper.py", line 47, in <module>
    main(argv=sys.argv[1:])
  File "/m2b/home/hechth/git/recetox/galaxytools/tools/msmetaenhancer/msmetaenhancer_wrapper.py", line 31, in main
    asyncio.run(app.annotate_spectra(services, jobs))
  File "/home/hechth/miniconda3/envs/__msmetaenhancer@0.1.1/lib/python3.8/asyncio/runners.py", line 44, in run
    return loop.run_until_complete(main)
  File "/home/hechth/miniconda3/envs/__msmetaenhancer@0.1.1/lib/python3.8/asyncio/base_events.py", line 616, in run_until_complete
    return future.result()
  File "/home/hechth/miniconda3/envs/__msmetaenhancer@0.1.1/lib/python3.8/site-packages/MSMetaEnhancer/app.py", line 83, in annotate_spectra
    jobs = convert_to_jobs(jobs)
  File "/home/hechth/miniconda3/envs/__msmetaenhancer@0.1.1/lib/python3.8/site-packages/MSMetaEnhancer/libs/utils/Job.py", line 28, in convert_to_jobs
    return [Job(data) for data in jobs]
  File "/home/hechth/miniconda3/envs/__msmetaenhancer@0.1.1/lib/python3.8/site-packages/MSMetaEnhancer/libs/utils/Job.py", line 28, in <listcomp>
    return [Job(data) for data in jobs]
  File "/home/hechth/miniconda3/envs/__msmetaenhancer@0.1.1/lib/python3.8/site-packages/MSMetaEnhancer/libs/utils/Job.py", line 6, in __init__
    self.source, self.target, self.service = data
ValueError: too many values to unpack (expected 3)

- and avoid using "jobs" in other than Galaxy context

hechth · 2021-12-10T13:20:22Z

Also, when running CTS: name -> inchikey, the job tuns for about 1sec and then terminates as green but no data is added - is CTS currently down?

martenson · 2021-12-10T13:23:10Z

I can confirm what helge reported above (ValueError: too many values to unpack (expected 3))

Additionally I've found a discrepancy with how the select field for the conversions without order behaves (tldr I think it does not like the format of value in the <option> tags).

Once you select an option from the offered list it is supposed to disappear from the list, hence if you select all of them the list should offer you nothing more. This is not the case with how it works for this tool and I assume that is because the select2 implementation has trouble comparing string values like this: value="('name', 'inchikey', 'PubChem')". I suggest we adjust the values here to preserve the original UX behavior, otherwise this could be confusing to users (and allows duplicates -- screenshot below).This adjustment may require additional mapping in our wrapper.

xtrojak · 2021-12-13T07:53:17Z

I found another bug when running only a single job (Pubchem: name -> inchikey)

Traceback (most recent call last):
  File "/m2b/home/hechth/git/recetox/galaxytools/tools/msmetaenhancer/msmetaenhancer_wrapper.py", line 47, in <module>
    main(argv=sys.argv[1:])
  File "/m2b/home/hechth/git/recetox/galaxytools/tools/msmetaenhancer/msmetaenhancer_wrapper.py", line 31, in main
    asyncio.run(app.annotate_spectra(services, jobs))
  File "/home/hechth/miniconda3/envs/__msmetaenhancer@0.1.1/lib/python3.8/asyncio/runners.py", line 44, in run
    return loop.run_until_complete(main)
  File "/home/hechth/miniconda3/envs/__msmetaenhancer@0.1.1/lib/python3.8/asyncio/base_events.py", line 616, in run_until_complete
    return future.result()
  File "/home/hechth/miniconda3/envs/__msmetaenhancer@0.1.1/lib/python3.8/site-packages/MSMetaEnhancer/app.py", line 83, in annotate_spectra
    jobs = convert_to_jobs(jobs)
  File "/home/hechth/miniconda3/envs/__msmetaenhancer@0.1.1/lib/python3.8/site-packages/MSMetaEnhancer/libs/utils/Job.py", line 28, in convert_to_jobs
    return [Job(data) for data in jobs]
  File "/home/hechth/miniconda3/envs/__msmetaenhancer@0.1.1/lib/python3.8/site-packages/MSMetaEnhancer/libs/utils/Job.py", line 28, in <listcomp>
    return [Job(data) for data in jobs]
  File "/home/hechth/miniconda3/envs/__msmetaenhancer@0.1.1/lib/python3.8/site-packages/MSMetaEnhancer/libs/utils/Job.py", line 6, in __init__
    self.source, self.target, self.service = data
ValueError: too many values to unpack (expected 3)

Thank you for noticing! Fixed in 906c767.

xtrojak · 2021-12-13T07:55:05Z

Additionally I've found a discrepancy with how the select field for the conversions without order behaves (tldr I think it does not like the format of value in the <option> tags).

Once you select an option from the offered list it is supposed to disappear from the list, hence if you select all of them the list should offer you nothing more. This is not the case with how it works for this tool and I assume that is because the select2 implementation has trouble comparing string values like this: value="('name', 'inchikey', 'PubChem')". I suggest we adjust the values here to preserve the original UX behavior, otherwise this could be confusing to users (and allows duplicates -- screenshot below).This adjustment may require additional mapping in our wrapper.

Yes the problem was indeed in the format of values. I changed the format to a simpler string in 906c767 and now it behaves as expected.

xtrojak · 2021-12-13T07:56:19Z

Also, when running CTS: name -> inchikey, the job tuns for about 1sec and then terminates as green but no data is added - is CTS currently down?

CTS is running, but there are some new issues. This particular problem was already detected and mentioned here. It is btw a reason why tests for MSMetaEnhancer are currently failing.

… without .msp extension from version 0.11.0

martenson · 2022-01-07T12:16:04Z

Output output_file:  different than expected, difference (using diff):
( /tmp/tmpzccq9rgksample_out.msp v. /tmp/tmpouwehnqesample_out.msp )
--- local_file
+++ history_data
@@ -4,10 +4,10 @@
 CASNO: 1333-74-0
 ID: 1
 COMMENT: NIST MS# 245692, Seq# M1
+INCHIKEY: UFHFLCQGNIYNRP-UHFFFAOYSA-N
 INCHI: InChI=1S/H2/h1H
-INCHIKEY: UFHFLCQGNIYNRP-UHFFFAOYSA-N
-IUPAC_NAME: molecular hydrogen
-SMILES: [HH]
+IUPAC_NAME: hydrogen monohydride
+SMILES: [H][H]
 NUM PEAKS: 2
 1.0	20.98
 2.0	999.0
@@ -18,9 +18,9 @@
 CASNO: 7782-39-0
 ID: 2
 COMMENT: NIST MS# 61316, Seq# M2
+INCHIKEY: UFHFLCQGNIYNRP-VVKOMZTBSA-N
 INCHI: InChI=1S/H2/h1H/i1+1D
-INCHIKEY: UFHFLCQGNIYNRP-VVKOMZTBSA-N
-SMILES: [HH]
+SMILES: [2H][2H]
********
*SNIP *
********
 13.0	85.92
@@ -67,10 +65,9 @@
 CASNO: 74-82-8
 ID: 5
 COMMENT: Any=100 ; NIST MS# 18809, Seq# R27
+INCHIKEY: VNWKTOKETHGBQD-UHFFFAOYSA-N
 INCHI: InChI=1S/CH4/h1H4
-INCHIKEY: VNWKTOKETHGBQD-UHFFFAOYSA-N
 IUPAC_NAME: methane
-SMILES: C
 NUM PEAKS: 6
 12.0	7.99
 13.0	28.97
@@ -85,10 +82,9 @@
 CASNO: 74-82-8
 ID: 6
 COMMENT: Any=100 ; NIST MS# 423924, Seq# R28
+INCHIKEY: VNWKTOKETHGBQD-UHFFFAOYSA-N
 INCHI: InChI=1S/CH4/h1H4
-INCHIKEY: VNWKTOKETHGBQD-UHFFFAOYSA-N
 IUPAC_NAME: methane
-SMILES: C
 NUM PEAKS: 6
 12.0	25.98
 13.0	74.93

xtrojak · 2022-01-07T15:07:19Z

Output output_file:  different than expected, difference (using diff):
( /tmp/tmpzccq9rgksample_out.msp v. /tmp/tmpouwehnqesample_out.msp )
--- local_file
+++ history_data
@@ -4,10 +4,10 @@
 CASNO: 1333-74-0
 ID: 1
 COMMENT: NIST MS# 245692, Seq# M1
+INCHIKEY: UFHFLCQGNIYNRP-UHFFFAOYSA-N
 INCHI: InChI=1S/H2/h1H
-INCHIKEY: UFHFLCQGNIYNRP-UHFFFAOYSA-N
-IUPAC_NAME: molecular hydrogen
-SMILES: [HH]
+IUPAC_NAME: hydrogen monohydride
+SMILES: [H][H]
 NUM PEAKS: 2
 1.0	20.98
 2.0	999.0
@@ -18,9 +18,9 @@
 CASNO: 7782-39-0
 ID: 2
 COMMENT: NIST MS# 61316, Seq# M2
+INCHIKEY: UFHFLCQGNIYNRP-VVKOMZTBSA-N
 INCHI: InChI=1S/H2/h1H/i1+1D
-INCHIKEY: UFHFLCQGNIYNRP-VVKOMZTBSA-N
-SMILES: [HH]
+SMILES: [2H][2H]
********
*SNIP *
********
 13.0	85.92
@@ -67,10 +65,9 @@
 CASNO: 74-82-8
 ID: 5
 COMMENT: Any=100 ; NIST MS# 18809, Seq# R27
+INCHIKEY: VNWKTOKETHGBQD-UHFFFAOYSA-N
 INCHI: InChI=1S/CH4/h1H4
-INCHIKEY: VNWKTOKETHGBQD-UHFFFAOYSA-N
 IUPAC_NAME: methane
-SMILES: C
 NUM PEAKS: 6
 12.0	7.99
 13.0	28.97
@@ -85,10 +82,9 @@
 CASNO: 74-82-8
 ID: 6
 COMMENT: Any=100 ; NIST MS# 423924, Seq# R28
+INCHIKEY: VNWKTOKETHGBQD-UHFFFAOYSA-N
 INCHI: InChI=1S/CH4/h1H4
-INCHIKEY: VNWKTOKETHGBQD-UHFFFAOYSA-N
 IUPAC_NAME: methane
-SMILES: C
 NUM PEAKS: 6
 12.0	25.98
 13.0	74.93

This is caused by recent changes in MSME regarding values validation. Could be easily fixed by updating the expected file.

However, we agreed that this tool cannot reproduce consistent results among different runs. All tests checking for data content were removed from MSME. Do we want to remove this test also from Galaxy wrapper? We probably can't rely on them, although they use quite simple molecules where metadata is expected to be rather consistent.

martenson · 2022-01-07T15:48:24Z

@xtrojak I guess you need to make a call whether the test is useful or not. My position is that if it breaks ~once a year, it could still be useful, if it is every month then the maintenance burden is too high.

hechth · 2022-01-07T16:20:58Z

I'd suggest staying with it for now and when we see that it becomes too maintenance-intensive we remove it.

hechth · 2022-01-10T08:50:18Z

Looks good to me, I think this can be merged. Thanks for the great work!

xtrojak · 2022-01-10T08:50:31Z

Alright, I have updated the test output file for now, we will see how it goes in the future.

martenson · 2022-01-13T12:38:44Z

Well done @xtrojak. The tool is now on UMSA here: https://umsa.cerit-sc.cz/root?tool_id=testtoolshed.g2.bx.psu.edu/repos/recetox/msmetaenhancer/msmetaenhancer/0.1.2+galaxy0

xtrojak added 6 commits November 22, 2021 09:42

Added simple wrapper (cannot choose jobs) for MSMetaEnhancer

d8a260a

Added initial version of MSMetaEnhancer tool config

b8ea502

Improved wrapper to fit format of passed services

1537f31

Fixed typos in shed

76895e4

Fixed problem with msp file extension required by matchms export

3ea60e9

Added simple test

2d39cc3

xtrojak requested review from martenson and hechth December 2, 2021 15:51

hechth requested changes Dec 3, 2021

View reviewed changes

Fixed test files

8569605

xtrojak marked this pull request as draft December 6, 2021 08:43

xtrojak added 2 commits December 6, 2021 12:55

Specify conversion job options in macro file

3ed7e73

Implemented specification of particular converion jobs

7429f24

Improved linting as requested by flake

9d3e282

xtrojak marked this pull request as ready for review December 6, 2021 12:24

martenson reviewed Dec 9, 2021

View reviewed changes

tools/msmetaenhancer/msmetaenhancer.xml Outdated Show resolved Hide resolved

tools/msmetaenhancer/msmetaenhancer.xml Outdated Show resolved Hide resolved

Fixed arguments order in job_options macro

24ae942

Minor tool xml config changes, switched to MSME 0.1.1

da914e9

Added instructions to update job_options in macro when updating the tool

549080a

clarify some language of the help

65f8d36

- and avoid using "jobs" in other than Galaxy context

Changed format of job options to work properly with multi select

906c767

xtrojak and others added 2 commits January 6, 2022 12:55

Removed creating of temp files because matchms allows export to files…

b875428

… without .msp extension from version 0.11.0

bump MSME version

b247b51

Updated test file to reflect the changes in MSMetaEnhancer 0.1.2

8308531

hechth approved these changes Jan 10, 2022

View reviewed changes

hechth merged commit 2c9c75f into RECETOX:master Jan 11, 2022

xtrojak mentioned this pull request Mar 21, 2022

upload all repos with green light to MTS #236

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add MSmetaEnhancer #199

Add MSmetaEnhancer #199

xtrojak commented Dec 2, 2021 •

edited

Loading

xtrojak commented Dec 6, 2021 •

edited

Loading

hechth commented Dec 9, 2021

martenson commented Dec 9, 2021

martenson left a comment •

edited

Loading

hechth commented Dec 9, 2021

hechth commented Dec 9, 2021

xtrojak commented Dec 9, 2021

xtrojak commented Dec 9, 2021

hechth commented Dec 10, 2021

hechth commented Dec 10, 2021

martenson commented Dec 10, 2021 •

edited

Loading

xtrojak commented Dec 13, 2021

xtrojak commented Dec 13, 2021

xtrojak commented Dec 13, 2021

martenson commented Jan 7, 2022

xtrojak commented Jan 7, 2022

martenson commented Jan 7, 2022

hechth commented Jan 7, 2022

hechth commented Jan 10, 2022

xtrojak commented Jan 10, 2022

martenson commented Jan 13, 2022

Add MSmetaEnhancer #199

Add MSmetaEnhancer #199

Conversation

xtrojak commented Dec 2, 2021 • edited Loading

xtrojak commented Dec 6, 2021 • edited Loading

hechth commented Dec 9, 2021

martenson commented Dec 9, 2021

martenson left a comment • edited Loading

Choose a reason for hiding this comment

hechth commented Dec 9, 2021

hechth commented Dec 9, 2021

xtrojak commented Dec 9, 2021

xtrojak commented Dec 9, 2021

hechth commented Dec 10, 2021

hechth commented Dec 10, 2021

martenson commented Dec 10, 2021 • edited Loading

xtrojak commented Dec 13, 2021

xtrojak commented Dec 13, 2021

xtrojak commented Dec 13, 2021

martenson commented Jan 7, 2022

xtrojak commented Jan 7, 2022

martenson commented Jan 7, 2022

hechth commented Jan 7, 2022

hechth commented Jan 10, 2022

xtrojak commented Jan 10, 2022

martenson commented Jan 13, 2022

xtrojak commented Dec 2, 2021 •

edited

Loading

xtrojak commented Dec 6, 2021 •

edited

Loading

martenson left a comment •

edited

Loading

martenson commented Dec 10, 2021 •

edited

Loading