Confirmed that the test script to run the example datapack included in the pipeline can be processed.

This is an example result for objId=55. 1D spectrum shows [OII], Hb, [OIII] at z=0.853 clearly, but DRP1D returns the best-fit z of 0.08. Needs to further investigation...

Hi Kiyoto,

Sorry to introduce inside your ticket. If you want any help on this spectrum, we'll be glad to contribuate.

I confirmed that it returns a good redshift result if we limit the redshift range very narrowly, like  

    "redshiftrange": [
        "0.8",
        "0.9"
    ],

Pierre-Yves CHABAUD Could you run drp1d for the spectrum below in your side? I also appreciate if you tell me a good parameter configuration?

pfsObject-001-00000-0,0-0000000000000037-002-0xe0241444a9af118c.fits

I observe a non physical high pick value at pixel 346. By reducing lambda range to

"lambdarange":["3800","13000"]

the pick is ignored and the right redshift appears in the candidates list. However, the right redshift remains in lower probability position. We are still investigating on this case.

Dear Kyoto,

I think I have identified the problem : there is a strong noise peak at 6330 A which is not identified in the variance. This ear is thus used as Ly-alpha, CIII and OII as first solution since it is very strong and has thus more weight than the true emission lines. 

Thank you for letting me know your results vlebrun & Pierre-Yves CHABAUD (I should have replied earlier). So, in this case, the best redshift is ~4, right? I could not reproduce the same result in my environment. Could you send me your configuration file in json format?

Hi Kiyoto Yabe , by looking at this issue, we've just found a bug in the pipeline on parameters file overloading ( https://pfspipe.ipmu.jp/jira/browse/PIPE1D-22 ). I propose to wait for fixing it, before going further. Sorry for the delay.

Pierre-Yves CHABAUD  OK, thank you for letting me know 

I'm investigating the feature at around 633 nm, which was reproduced with the latest 2D pipeline (`w.2020.39`). As shown in the plots below, this seems to come out after making `pfsMerged` spectrum in only `visit=24`. The plots are for only 2 objects (objId=0x37 and 0x12), but I see the same feature for other fibers. price do you have any idea?

I am investigating (PIPE2D-633; the number of the ticket is completely coincidental!).

This is the same object as before but with a fix of PIPE2D-635

The 1D output looks reasonable.

good to see we are able to measure the redshift on such strong emission lines...

I just confirmed that the processing is successfully done with the master branch after PIPE2D-635 is merged. So, I will close this ticket.