ocbrdf M02 IndexError

[mikemaniscalco]

I am having trouble with the running oc_brdf with brdf_model = "M02". Other models such as L11 run fine.
Potentially related to:  a550866I I run oc_brdf.brdf_prototype(ds, brdf_model='M02') I get the below error message.

I might be thinking about this wrong, but it seems like modifying lines 149-153 of brdf_model_M02.py to index fQ_chl similar to Rrs and f0_chl might make sense?:
R442 = np.pi * Rrs.sel(bands=b442) * f0_chl.sel(bands_f0=2.0) / fQ_chl.sel(bands=b442).interp(wavelengths_FOQ=b442)
Original:
R442 = np.pi * Rrs.sel(bands=b442) * f0_chl.sel(bands_f0=2.0) / fQ_chl.interp(wavelengths_FOQ=b442)

IndexError                                Traceback (most recent call last)
Cell In[3], line 23
     19 df = df.set_index(['bands'])
     21 ds = xr.Dataset.from_dataframe(df)
---> 23 oc_brdf.brdf_prototype(ds, brdf_model='M02')

File ~/Documents/waterhypernet/code/ocbrdf/ocbrdf_main.py:81, in brdf_prototype(ds, adf, brdf_model)
     78     ds['log10_chl'] = 0 * ds['sza'] + float(np.log10(BRDF_model.OC4MEchl0))
     79     chl_iter[-1]    = 0 * ds['sza'] + float(BRDF_model.OC4MEchl0)
---> 81 ds = BRDF_model.backward(ds, iter_brdf)
     83 # M02: Check convergence (dummy for M02SeaDAS for the moment... epsilon set to 0)
     84 if brdf_model in ['M02', 'M02SeaDAS']:

File ~/Documents/waterhypernet/code/ocbrdf/brdf_model_M02.py:149, in M02.backward(self, ds, iter_brdf)
    146 fQ_chl = drop_unused_coords(fQ_chl)
    148 # Get Rrs at relevant bands for OC4ME and convert to Irradiance Reflectance (R)
--> 149 R442 = np.pi * Rrs.sel(bands=b442) * f0_chl.sel(bands_f0=2.0) / fQ_chl.interp(wavelengths_FOQ=b442)
    150 R490 = np.pi * Rrs.sel(bands=b490) * f0_chl.sel(bands_f0=3.0) / fQ_chl.interp(wavelengths_FOQ=b490)
    151 R510 = np.pi * Rrs.sel(bands=b510) * f0_chl.sel(bands_f0=4.0) / fQ_chl.interp(wavelengths_FOQ=b510)

File ~/miniforge3/lib/python3.10/site-packages/xarray/core/dataarray.py:2204, in DataArray.interp(self, coords, method, assume_sorted, kwargs, **coords_kwargs)
   2199 if self.dtype.kind not in "uifc":
   2200     raise TypeError(
   2201         "interp only works for a numeric type array. "
   2202         "Given {}.".format(self.dtype)
   2203     )
-> 2204 ds = self._to_temp_dataset().interp(
   2205     coords,
   2206     method=method,
   2207     kwargs=kwargs,
   2208     assume_sorted=assume_sorted,
   2209     **coords_kwargs,
   2210 )
   2211 return self._from_temp_dataset(ds)

File ~/miniforge3/lib/python3.10/site-packages/xarray/core/dataset.py:3515, in Dataset.interp(self, coords, method, assume_sorted, kwargs, method_non_numeric, **coords_kwargs)
   3512         variables[k] = v
   3514 # Extract coordinates from indexers
-> 3515 coord_vars, new_indexes = selected._get_indexers_coords_and_indexes(coords)
   3516 variables.update(coord_vars)
   3517 indexes.update(new_indexes)

File ~/miniforge3/lib/python3.10/site-packages/xarray/core/dataset.py:2443, in Dataset._get_indexers_coords_and_indexes(self, indexers)
   2439 # we don't need to call align() explicitly or check indexes for
   2440 # alignment, because merge_variables already checks for exact alignment
   2441 # between dimension coordinates
   2442 coords, indexes = merge_coordinates_without_align(coords_list)
-> 2443 assert_coordinate_consistent(self, coords)
   2445 # silently drop the conflicted variables.
   2446 attached_coords = {k: v for k, v in coords.items() if k not in self._variables}

File ~/miniforge3/lib/python3.10/site-packages/xarray/core/coordinates.py:476, in assert_coordinate_consistent(obj, coords)
    473 for k in obj.dims:
    474     # make sure there are no conflict in dimension coordinates
    475     if k in coords and k in obj.coords and not coords[k].equals(obj[k].variable):
--> 476         raise IndexError(
    477             f"dimension coordinate {k!r} conflicts between "
    478             f"indexed and indexing objects:\n{obj[k]}\nvs.\n{coords[k]}"
    479         )

IndexError: dimension coordinate 'bands' conflicts between indexed and indexing objects:
<xarray.DataArray 'bands' (bands: 161)>
array([400. , 402.5, 405. , 407.5, 410. , 412.5, 415. , 417.5, 420. , 422.5,
       425. , 427.5, 430. , 432.5, 435. , 437.5, 440. , 442.5, 445. , 447.5,
       450. , 452.5, 455. , 457.5, 460. , 462.5, 465. , 467.5, 470. , 472.5,
       475. , 477.5, 480. , 482.5, 485. , 487.5, 490. , 492.5, 495. , 497.5,
       500. , 502.5, 505. , 507.5, 510. , 512.5, 515. , 517.5, 520. , 522.5,
       525. , 527.5, 530. , 532.5, 535. , 537.5, 540. , 542.5, 545. , 547.5,
       550. , 552.5, 555. , 557.5, 560. , 562.5, 565. , 567.5, 570. , 572.5,
       575. , 577.5, 580. , 582.5, 585. , 587.5, 590. , 592.5, 595. , 597.5,
       600. , 602.5, 605. , 607.5, 610. , 612.5, 615. , 617.5, 620. , 622.5,
       625. , 627.5, 630. , 632.5, 635. , 637.5, 640. , 642.5, 645. , 647.5,
       650. , 652.5, 655. , 657.5, 660. , 662.5, 665. , 667.5, 670. , 672.5,
       675. , 677.5, 680. , 682.5, 685. , 687.5, 690. , 692.5, 695. , 697.5,
       700. , 702.5, 705. , 707.5, 710. , 712.5, 715. , 717.5, 720. , 722.5,
       725. , 727.5, 730. , 732.5, 735. , 737.5, 740. , 742.5, 745. , 747.5,
       750. , 752.5, 755. , 757.5, 760. , 762.5, 765. , 767.5, 770. , 772.5,
       775. , 777.5, 780. , 782.5, 785. , 787.5, 790. , 792.5, 795. , 797.5,
       800. ])
Coordinates:
  * bands    (bands) float64 400.0 402.5 405.0 407.5 ... 792.5 795.0 797.5 800.0
vs.
<xarray.Variable ()>
array(442.5)

[oceancolorcoder]

@mikemaniscalco I'm on PTO but hopefully Juan, who is most familiar with this code, will get back to you while I'm away.

[juanchossn]

Dear @mikemaniscalco, I think this issue is related to different ways of handling 1D np. arrays by x array .interp and .sel funcitons. It should be solved after PR #278, could you test? In my systems things were working fine before.

[mikemaniscalco]

@juanchossn That looks like it does the trick! thanks!

[oceancolorcoder]

@juanchossn Thanks for following up. Interesting that @mikemaniscalco was able to successfully tag you but when I try it doesn't want to link. This is a test for whether just typing it without it linking works anyway...

As we have resolution, I will close this discussion.