That is in the GFM file, though encoded in a slightly non-obvious way.

price How should we do this? Adding an extra column to the DESIGN/CONFIG tables would probably be best, but we'll have to think about backwards compatibility.
There was also talk about adding a relative transmission for each fibre (e.g. 3 is c. 10% due to a Cable A problem), should we do that at the same time? Currently there's just a binary decision "Is fibre BLOCKED?"

I think we need it in the pfsConfig as it looks as if e.g. we may need this information for sky subtraction.

I don't see the need to have it in the pfsConfig if it's in pfs_utils (pfs.utils.fiberids.FiberIds) and static. But if it's not static, then we need a way of minimally choosing the correct configuration.

Having the relative transmission might be nice (though it's theoretically available through the fiberProfiles), but again I don't see the need to put it in the pfsConfig.

Currently the pipelines don't use anything accessed via pfs_utils, and I think it makes things more complicated if we have to start doing so. My concern is what we should do if MTPs move together (and I think that there's evidence for that).
If it's just Fabrice in his lab, then I agree that documenting the interface for him would probably be enough.

We already distribute pfs_utils along with the 2D pipeline, and we use it to create pfsDesign files for DCB (makeDummyCableBDesign.py).

I think we agreed in a recent meeting that using the mapping in pfs_utils is sufficient. I propose closing this WONTFIX.

well

if so can we be sure that we have the proper API with pfs_utils and so a proper ticket? hassan ?

OK. Let me check. If there is missing/failing functionality I will file a new ticket.

I'm working on this (that is, I wrote the code). I assume that the name we want is e.g. D1-3-1

D1-3 tell us which mtp but not which fiber in the mtp connector, if I am correct

D1-3-<specId>-<mtp position>
it would be nice to have also the position in case we identify fiber that show bad illumination (as we did on SM1)

There are 4 of values per fibre (A, BA, BC, and C).  So do you want me to return

("D0-1-1", (2, 26, 31, 7))

or something else?

Or just the 4 raw strings, including the cobra number?

it may be worth to have the complete information, no ?
or we could have different option that returns different information ?

for my perspective I only need which mtp and position on Cable A but... so D0-1-x-26 for example or simply DO-1 according what I am looking for (mtp only for this last example)

As discussed during the DRP telecon on 2021-06-04, this functionality will be implemented in pfs_utils, not the datamodel. INSTRM-1288 covers that work.

Return MTP information for the specified fiberIds

        Args
        ----
        fiberIds : array of 1-indexed fiberIds
        pfsConfig : `pfs.datamodel.PfsConfig`
           Tell us e.g. which fibres go to SuNSS

        Returns
        -------
        an array of ("MTPID", (holes), cobraId) where "holes" are for the A, BA, BC, and C connectors,
        and cobraId is the 0-indexed global cobra ID

        For SuNSS, the "cobra ID" is negative, and its absolute value is the ID in the
        ferrule (for the imaging leg) or 127 + ID (for the diffuse leg)

The usage is:

import lsst.daf.persistence as dafPersist
import pfs.utils.fiberids as fiberids

butler = dafPersist.Butler("/projects/HSC/PFS/Subaru")
dataId = dict(visit=45783, arm="r", spectrograph=1)
pfsConfig = butler.get('pfsConfig', dataId)

fiberIds = fiberids.FiberIds()

fiberIds.fiberIdToMTP(pfsConfig.fiberId, pfsConfig)[100] 

which returns

('U1-2-1', [26, 2, 7, 31], -51)

(where -51 indicated an imaging SuNSS "cobra").

Pushed to tickets/DAMD-110

Proposed changes in branch look fine.

Merged to master.