Fix getting ResultsRange field

[Bugerman58]

Description of the issue/feature this PR addresses

by @toropok description

With DataBox report based on Analysis query we noticed that when DefaultResultsRangeProvider being called following exception raised immediately:

ERROR Archetypes None
Traceback (most recent call last):
  File "/home/senaite/buildout-cache/eggs/cp27mu/Products.Archetypes-1.16.6-py2.7.egg/Products/Archetypes/BaseObject.py", line 146, in initializeArchetype
    self.setDefaults()
  File "/home/senaite/buildout-cache/eggs/cp27mu/Products.Archetypes-1.16.6-py2.7.egg/Products/Archetypes/BaseObject.py", line 445, in setDefaults
    self.Schema().setDefaults(self)
  File "/home/senaite/buildout-cache/eggs/cp27mu/Products.Archetypes-1.16.6-py2.7.egg/Products/Archetypes/Schema/__init__.py", line 504, in setDefaults
    default = field.getDefault(instance)
  File "/home/senaite/buildout-cache/eggs/cp27mu/Products.Archetypes-1.16.6-py2.7.egg/Products/Archetypes/Field.py", line 693, in getDefault
    return default_adapter()
  File "/home/senaite/senaitelims/src/senaite.core/src/bika/lims/browser/fields/resultrangefield.py", line 104, in __call__
    field = sample.getField("ResultsRange")
AttributeError: 'RequestContainer' object has no attribute 'getField'

Investigation shown that it comes from resultrangefield.py line 103: field = sample.getField("ResultsRange")

The DefaultResultsRangeProvider docstring states that it follows backward compatibility solution for PR: #1506

I've added couple hasattr checks to exclude exceptions riot in logs.

Current behavior before PR

Exception thrown

Desired behavior after PR is merged

tomb silence in logs 🙂 everything goes nice and easy

--
I confirm I have tested this PR thoroughly and coded it according to PEP8
and Plone's Python styleguide standards.

[ramonski]

Thanks @Bugerman58 for the fixture.
Although I don't get fully why this happens, I would suggest some changes to make the whole thing more readable/understandable.
Maybe you can log an informative message or add a comment why the attribute might be missing?

[ramonski]

Please refactor to a class property as follows:

@property
def specification(self):
    spec = None
    try:
        spec = self.analysisrequest.getSpecification()
    except AttributeError:
        pass # or maybe log an error?
    return spec

[ramonski]

Please refactor as follows:

    ...
    rr = self.get_results_range_for(sample)

def get_results_range(self, sample):
    rr = None
    try:
        sample.getResultsRange()
    except AttributeError:
        pass # or log an error?
    return rr

[Bugerman58]

Thanks @Bugerman58 for the fixture. Although I don't get fully why this happens, I would suggest some changes to make the whole thing more readable/understandable. Maybe you can log an informative message or add a comment why the attribute might be missing?

This exception arises of creating report for 'Anaylsis' content type and every time when update query for this report.

To find out the reason for this exception, we added logging for this object and found:

...
'<Products.ATContentTypes.tool.factory.TempFolder object at 0x7f6e3c3f5b50>': 'RequestContainer' object has no attribute 'getField'
...

, and after that:

...
'<Products.ATContentTypes.tool.factory.TempFolder object at 0x7f6e3c3f5b50>': 'RequestContainer' object has no attribute 'getSpecification'
...

[ramonski]

Please always use the field getter instead of accessing the attribute directly.

Maybe you can simplify the whole thing with a combination this:

 def get_results_range_for(self, sample, analysis):
    rr = {}
    try:
        field = sample.getField("ResultsRange", {})
        rr = field.get(sample, search_by=analysis.getKeyword())
     except AttributeError:
        pass
     return rr

Also use a more descriptive logging message. As you know that it is a sample, you could be more specific, e.g.

logger.error("Failed to get results range for sample '{}' and analysis '{}': "
             "{}".format(sample.getId(), analysis.getKeyword(), str(e)))

[Bugerman58]

So, if at look on this code:

# Try with uid (this shouldn't be necessary)
service_uid = analysis.getServiceUID()

, getting the results range is solved by checking the existence of the Keyword. Right?

Or to get results range by UID has the meaning?

[ramonski]

Good question. I'm not sure why we added this fallback. But maybe you can bake it in for backwards compatibility?

[Bugerman58]

sorry for my carelessness (a large of commits)

[ramonski]

Great, thanks @Bugerman58. Much cleaner now!