geospatial ground irradiance

pvdeg/geospatial.py

@@ -7,8 +7,10 @@
     humidity,
     letid,
     utilities,
+    pysam,
 )
 
+
 import xarray as xr
 import dask.array as da
 import pandas as pd
@@ -27,6 +29,8 @@
 from typing import Tuple
 from shapely import LineString, MultiLineString
 
+# {var: (dim, da.empty([dims_size[d] for d in dim]), attrs.get(var)) for var, dim in shapes.items()}
+
 def start_dask(hpc=None):
     """
     Starts a dask cluster for parallel processing.
@@ -88,8 +92,31 @@ def start_dask(hpc=None):
 
     return client
 
+def _ds_from_arbitrary(res, func):
+    """
+    Convert an arbitrary return type to xarray.Dataset. 
+    """
+
+    if isinstance(res, pysam.inspirePysamReturn):
+        return pysam._handle_pysam_return(res)
+    # add more conditionals if we have special cases 
+    # or add general case for mixed return dimensions: HARD
+
+    # handles collections with elements of same shapes
+    df = _df_from_arbitrary(res=res, func=func)
+    ds = xr.Dataset.from_dataframe(df)
+
+    if not df.index.name:
+        ds = ds.isel(index=0, drop=True)
+
+    return ds
+
 
 def _df_from_arbitrary(res, func):
+    """
+    Convert an arbitrary return type to dataframe.
+    Results must be of similar shape currently. Either all numerics or all timeseries.
+    """
     numerics = (int, float, np.number)
     arrays = (np.ndarray, pd.Series)
 
@@ -102,12 +129,13 @@ def _df_from_arbitrary(res, func):
     elif isinstance(res, tuple) and all(isinstance(item, numerics) for item in res):
         return pd.DataFrame([res])
     elif isinstance(res, tuple) and all(isinstance(item, arrays) for item in res):
+        # add check for same size, raise value error otherwise
         return pd.concat(
             res, axis=1
         )  # they must all be the same length here or this will error out
     else:
         raise NotImplementedError(
-            f"function return type: {type(res)} not available for geospatial analysis yet."
+            f"function return type: {type(res)} not available for geospatial analysis yet. This could be result of mismatched coordinates of outputs. EX. tuple(dataframe, int)."
         )
 
 
@@ -137,20 +165,15 @@ def calc_gid(ds_gid, meta_gid, func, **kwargs):
     if type(meta_gid["latitude"]) == dict:
         meta_gid = utilities.fix_metadata(meta_gid)
 
-    df_weather = (
-        ds_gid.to_dataframe()
-    )  # set time index here? is there any reason the weather shouldn't always have only pd.datetime index? @ martin?
+    # set time index here? is there any reason the weather shouldn't always have only pd.datetime index? @ martin?
+    df_weather = ds_gid.to_dataframe()  
     if isinstance(
         df_weather.index, pd.MultiIndex
     ):  # check for multiindex and convert to just time index, don't know what was causing this
         df_weather = df_weather.reset_index().set_index("time")
 
     res = func(weather_df=df_weather, meta=meta_gid, **kwargs)
-    df_res = _df_from_arbitrary(res, func)  # convert all return types to dataframe
-    ds_res = xr.Dataset.from_dataframe(df_res)
-
-    if not df_res.index.name:
-        ds_res = ds_res.isel(index=0, drop=True)
+    ds_res = _ds_from_arbitrary(res, func)
 
     return ds_res
 
@@ -268,17 +291,39 @@ def output_template(
     dims = set([d for dim in shapes.values() for d in dim])
     dims_size = dict(ds_gids.sizes) | add_dims
 
+    # we need to properly add the dimensions
+    # can we redo with dictionary comprehension
+    # best approach @martin?
+    coords = {}
+    for dim in dims:
+        if dim in ds_gids:
+            coords[dim] = ds_gids[dim]
+        elif dim in add_dims:
+            coords[dim] = np.arange(dims_size[dim]) # i dont like this
+        else:
+            raise ValueError(f"dim: {dim} not in ds_gids or add_dims")
+
+
     output_template = xr.Dataset(
         data_vars={
             var: (dim, da.empty([dims_size[d] for d in dim]), attrs.get(var)) # this will produce a dask array with 1 chunk of the same size as the input
             for var, dim in shapes.items()
         },
-        coords={dim: ds_gids[dim] for dim in dims},
+        # moved as part of the above changes
+        # coords={dim: ds_gids[dim] for dim in dims},
+        coords=coords,
         attrs=global_attrs,
     ) 
-
-    if ds_gids.chunks: # chunk to match input
-        output_template = output_template.chunk({dim: ds_gids.chunks[dim] for dim in dims})
+
+    # chunk to match input
+    if ds_gids.chunks: 
+        output_template = output_template.chunk(
+            {   
+                dim: ds_gids.chunks[dim] 
+                for dim in dims
+                if dim in ds_gids
+            }   # only chunk dimensions existing in the input
+        )
 
     return output_template
 

pvdeg/pysam.py

@@ -1,5 +1,11 @@
+""" 
+Pysam Integration for pvdeg, supports single site and geospatial calculations.
+Produced to support Inspire Agrivoltaics: https://openei.org/wiki/InSPIRE
+"""
+
 import dask.dataframe as dd
 import pandas as pd
+import pickle
 import json
 import sys
 import os
@@ -14,6 +20,7 @@
 from pvdeg import (
     weather,
     utilities,
+    DATA_DIR
 )
 
 def vertical_POA(
@@ -316,7 +323,7 @@ def pysam(
 
     # Duplicate Columns in the dataframe seem to cause this issue
     # Error (-4) converting nested tuple 0 into row in matrix.
-    pysam_model.SolarResource.solar_resource_data = solar_resource 
+    pysam_model.SolarResource.solar_resource_data = sr 
     pysam_model.execute()
     outputs = pysam_model.Outputs.export()
 
@@ -359,11 +366,51 @@ def pysam_hourly_trivial(weather_df, meta):
 
     return outputs
 
+# required to safely unpack results during geospatial.analysis
+class inspirePysamReturn():
+    def __init__(self, annual_poa, ground_irradiance, timeseries_index):
+        self.annual_poa = annual_poa
+        self.ground_irradiance = ground_irradiance
+        self.timeseries_index = timeseries_index
+
+
+# rename?
+import xarray as xr
+def _handle_pysam_return(pysam_res : inspirePysamReturn) -> xr.Dataset:
+    """Handle a pysam return object and transform it to an xarray"""
+
+    import dask.array as da
+    import numpy as np
+
+    # redo this using numba?
+    ground_irradiance_values = da.from_array([row[1:] for row in pysam_res.ground_irradiance[1:]])
+
+    single_location_ds = xr.Dataset(
+        data_vars={
+            "annual_poa" : pysam_res.annual_poa, # scalar variable
+            "ground_irradiance" : (("time", "distance"), ground_irradiance_values)
+        },
+        coords={
+            # "time" : np.arange(17520), # this will probably break because of the values assigned here, should be a pd.datetimeindex instead
+            "time" : pysam_res.timeseries_index,
+            # "distance" : np.array(pysam_res.ground_irradiance[0][1:])
+            "distance" : np.arange(10), # this matches the dimension axis of the output_temlate dataset
+        }
+    ) 
+
+    return single_location_ds
+
+
 # annual_poa_nom, annual_poa_front, annual_poa_rear, poa_nom, poa_front, or poa_rear
-# TODO: add config file
+# TODO: add config file, multiple config files.
 def inspire_ground_irradiance(weather_df, meta):
     """
     Get ground irradiance array and annual poa irradiance for a given point using pvsamv1
+
+    Returns
+    --------
+    result : inspirePysamReturn
+        returns an custom class object so we can unpack it later.
     """
 
     sr = solar_resource_dict(weather_df=weather_df, meta=meta)
@@ -373,13 +420,25 @@ def inspire_ground_irradiance(weather_df, meta):
     model.execute()
     outputs = model.Outputs.export()
 
-    # we only want these two
+    outputs = pysam(
+        weather_df = weather_df,
+        meta = meta,
+        pv_model = "pysamv1",
+        pv_model_default = "FlatPlatePVCommercial", # should use config file instead
+        results = ["subarray1_ground_rear_spatial", "annual_poa_front"],
+    )
+
+    # these will be of very different shapes
+    # annual_poa_front is a single numeric
+    # subarray1_ground_rear_spatial is a 2d result
 
-    inspire_res = {}
-    inspire_res["subarray1_ground_rear_spatial"] = outputs["subarray1_ground_rear_spatial"]
-    inspire_res["annual_poa_front"] = outputs["annual_poa_front"]
+    result = inspirePysamReturn(
+            ground_irradiance = outputs["subarray1_ground_rear_spatial"],
+            annual_poa = outputs["annual_poa_front"],
+            timeseries_index=weather_df.index,
+        )
 
-    return inspire_res
+    return result
 
 def solar_resource_dict(weather_df, meta):
     """
@@ -409,3 +468,10 @@ def solar_resource_dict(weather_df, meta):
     }
 
     return sr 
+
+def sample_pysam_result(weather_df, meta): # throw weather, meta away
+    """returns a sample inspirePysamReturn"""
+    with open(os.path.join(DATA_DIR,"inspireInstance.pkl"), "rb") as file:
+        inspireInstance = pickle.load(file)
+
+    return inspireInstance

pvdeg/temperature.py

@@ -24,23 +24,23 @@ def map_model(temp_model: str, cell_or_mod: str) -> callable:
 
     # double check that models are in correct maps
     module = {  # only module
-        "sapm": pvlib.temperature.sapm_module,
-        "sapm_mod": pvlib.temperature.sapm_module,
+        "sapm"          : pvlib.temperature.sapm_module,
+        "sapm_mod"      : pvlib.temperature.sapm_module,
     }
 
     cell = {  # only cell
-        "sapm": pvlib.temperature.sapm_cell,
-        "sapm_cell": pvlib.temperature.sapm_cell,
-        "pvsyst": pvlib.temperature.pvsyst_cell,
-        "ross": pvlib.temperature.ross,
-        "noct_sam": pvlib.temperature.noct_sam,
+        "sapm"          : pvlib.temperature.sapm_cell,
+        "sapm_cell"     : pvlib.temperature.sapm_cell,
+        "pvsyst"        : pvlib.temperature.pvsyst_cell,
+        "ross"          : pvlib.temperature.ross,
+        "noct_sam"      : pvlib.temperature.noct_sam,
         "generic_linear": pvlib.temperature.generic_linear,
     }
 
     agnostic = {  # module or cell
-        "faiman": pvlib.temperature.faiman,
-        "faiman_rad": pvlib.temperature.faiman_rad,
-        "fuentes": pvlib.temperature.fuentes,
+        "faiman"        : pvlib.temperature.faiman,
+        "faiman_rad"    : pvlib.temperature.faiman_rad,
+        "fuentes"       : pvlib.temperature.fuentes,
     }
 
     super_map = {"module": module, "cell": cell}
@@ -409,3 +409,23 @@ def temperature(
     temperature = func(**model_args)
 
     return temperature
+
+def _mixed_res(weather_df, meta):
+    """
+    geospatial test function. returns have mixed dimensions. the first is a timeseries, the second is a float.
+    This function is meant to discover problems with geospatial.analysis and its subroutines.
+
+    .. code_block : Python 
+
+        Shapes = {
+            "temperatures" : ("gid", "time"),
+            "avg_temp" : ("gid", ),
+        }
+    """
+
+    timeseries_df = pd.DataFrame(pvdeg.temperature.module(weather_df, meta))
+    avg_temp = timeseries_df[0].mean()
+
+    # now we have this problem. how can we run this
+    # return {'temperatures' : timeseries_df, 'avg_temp' : avg_temp}
+    return timeseries_df, avg_temp

tutorials_and_tools/tutorials_and_tools/PYSam.ipynb

@@ -168,9 +168,21 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 1,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "ename": "NameError",
+     "evalue": "name 'pvdeg' is not defined",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
+      "Cell \u001b[0;32mIn[1], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m location_grabber \u001b[38;5;241m=\u001b[39m pvdeg\u001b[38;5;241m.\u001b[39mscenario\u001b[38;5;241m.\u001b[39mGeospatialScenario()\n\u001b[1;32m      3\u001b[0m location_grabber\u001b[38;5;241m.\u001b[39maddLocation(country\u001b[38;5;241m=\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mUnited States\u001b[39m\u001b[38;5;124m\"\u001b[39m, downsample_factor\u001b[38;5;241m=\u001b[39m\u001b[38;5;241m80\u001b[39m)\n",
+      "\u001b[0;31mNameError\u001b[0m: name 'pvdeg' is not defined"
+     ]
+    }
+   ],
    "source": [
     "location_grabber = pvdeg.scenario.GeospatialScenario()\n",
     "\n",
@@ -228,9 +240,9 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "deg",
+   "display_name": "rpp",
    "language": "python",
-   "name": "python3"
+   "name": "rpp"
   },
   "language_info": {
    "codemirror_mode": {
@@ -242,9 +254,9 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.9.19"
+   "version": "3.10.14"
   }
  },
  "nbformat": 4,
- "nbformat_minor": 2
+ "nbformat_minor": 4
 }