grabbing results without stored scoped-key.dat

[JenkeScheen]

I've run some repex on a moonshot ligand series using Irfan's example; during my setup+submission I did not write out a scoped-key.dat, so I'm trying to retrieve this so I can get per-edge results.

Minimal code:

# [imports]
user_id = 'jscheen'
user_key = getpass()

asc = AlchemiscaleClient('https://api.alchemiscale.org', user_id, user_key)

scope = Scope('asap', 'jscheen', 'test2_moonshot_cpds')

network = asc.get_network(asc.query_networks()[0]) # this is not ideal; would be easier if we can query using scope?

This appears to work just fine as-is. But when I try to grab per-edge results:

for transformation in network.edges:
    sk = asc.get_scoped_key(transformation, scope)

The following exception gets thrown:

---------------------------------------------------------------------------
AlchemiscaleClientError                   Traceback (most recent call last)
Cell In[23], line 5
      3 for transformation in network.edges:
      4     sk = asc.get_scoped_key(transformation, scope)
----> 5     result = asc.get_transformation_results(sk)
      6     runtype = scan_components(transformation.stateA)
      7     mapping = transformation.mapping['ligand']

File ~/miniconda3/envs/alchemiscale-client/lib/python3.9/site-packages/alchemiscale/interface/client.py:449, in AlchemiscaleClient.get_transformation_results(self, transformation, return_protocoldagresults)
    425 """Get a `ProtocolResult` for the given `Transformation`.
    426 
    427 A `ProtocolResult` object corresponding to the `Protocol` used for this
   (...)
    444 
    445 """
    447 if not return_protocoldagresults:
    448     # get the transformation if we intend to return a ProtocolResult
--> 449     tf: Transformation = self.get_transformation(transformation)
    451 # get all protocoldagresultrefs for the given transformation
    452 protocoldagresultrefs = self._get_resource(
    453     f"/transformations/{transformation}/results",
    454 )

File ~/miniconda3/envs/alchemiscale-client/lib/python3.9/site-packages/alchemiscale/interface/client.py:110, in AlchemiscaleClient.get_transformation(self, transformation)
    107 def get_transformation(
    108     self, transformation: Union[ScopedKey, str]
    109 ) -> Transformation:
--> 110     return json_to_gufe(self._get_resource(f"/transformations/{transformation}"))

File ~/miniconda3/envs/alchemiscale-client/lib/python3.9/site-packages/alchemiscale/base/client.py:108, in AlchemiscaleBaseClient._retry.<locals>._wrapper(self, *args, **kwargs)
    106 while True:
    107     try:
--> 108         return f(self, *args, **kwargs)
    109     except (self._exception, AlchemiscaleConnectionError) as e:
    110         # if we are getting back HTTP errors and the status code is not
    111         # one of those we want to retry on, just raise
    112         if isinstance(e, self._exception):

File ~/miniconda3/envs/alchemiscale-client/lib/python3.9/site-packages/alchemiscale/base/client.py:164, in AlchemiscaleBaseClient._use_token.<locals>._wrapper(self, *args, **kwargs)
    159 # execute our function
    160 # if we get an unauthorized exception, it may be that our token is
    161 # stale; get a new one if so
    162 # if it's any other status code, raise it
    163 try:
--> 164     return f(self, *args, **kwargs)
    165 except self._exception as e:
    166     if e.status_code == 401:

File ~/miniconda3/envs/alchemiscale-client/lib/python3.9/site-packages/alchemiscale/base/client.py:212, in AlchemiscaleBaseClient._get_resource(self, resource, params)
    209     raise AlchemiscaleConnectionError(*e.args)
    211 if not 200 <= resp.status_code < 300:
--> 212     raise self._exception(
    213         f"Status Code {resp.status_code} : {resp.reason} : {resp.text}",
    214         status_code=resp.status_code,
    215     )
    216 content = json.loads(resp.text, cls=JSON_HANDLER.decoder)
    217 return content

AlchemiscaleClientError: Status Code 500 : Internal Server Error : Internal Server Error

.. which isn't very instructive. I think @dotsdl mentioned some of my edges failed - this error gets thrown on all of them, though? @dotsdl thoughts on how to proceed?

[ijpulidos]

Tagging @IAlibay and @dwhswenson from the openfe team that could probably benefit from the feedback here.

[dotsdl]

@JenkeScheen working on this now; will post conclusions here. I'm seeing a mismatch between the gufe keys for each object in this network in the DB and what I get client-side.

Can you tell me the versions of gufe and openfe that you used in your environment for submission?

[JenkeScheen]

gufe: 0.7.2
openfe: 0.7.3

[JenkeScheen]

update: just re-submitted a network with different versioning:
gufe=0.7.3
openfe=0.7.4