Seasonal and value index creation methods added

ecoscope/analysis/seasons.py

@@ -3,7 +3,6 @@
 import ee
 import numpy as np
 import pandas
-import pandas as pd
 import shapely
 import sklearn.mixture
 from scipy.stats import norm
@@ -36,9 +35,9 @@ def std_ndvi_vals(aoi=None, img_coll=None, band=None, img_scale=1, start=None, e
 
     ndvi_vals = coll.toBands().reduceRegion("mean", geo, bestEffort=True).values().getInfo()
 
-    df = pd.DataFrame(
+    df = pandas.DataFrame(
         {
-            "img_date": pd.to_datetime(coll.aggregate_array("system:time_start").getInfo(), unit="ms", utc=True),
+            "img_date": pandas.to_datetime(coll.aggregate_array("system:time_start").getInfo(), unit="ms", utc=True),
             "NDVI": ndvi_vals,
         }
     ).dropna(axis=0)
@@ -102,9 +101,41 @@ def seasonal_windows(ndvi_vals, cuts, season_labels):
         {
             "start": pandas.Series([grp["img_date"].iloc[0] for _, grp in grpd]),
             "end": pandas.Series([grp["img_date"].iloc[-1] for _, grp in grpd]).apply(
-                lambda x: x + pd.Timedelta(days=1)
+                lambda x: x + pandas.Timedelta(days=1)
             ),
             "season": pandas.Series(grp["season"].iloc[0] for name, grp in grpd),
             "unique_season": pandas.Series([name for name, _ in grpd]),
         }
     ).set_index("unique_season")
+
+
+def add_seasonal_index(df, index_name, start_date, end_date, aoi_geom_filter=None, seasons=2, season_labels=["dry", "wet"]):
+
+    aoi_ = None
+    try:
+        aoi_ = aoi_geom_filter.dissolve().iloc[0]['geometry']
+    except:
+        aoi_ = aoi_geom_filter
+
+    if len(season_labels) != seasons:
+        raise Exception(f"Parameter value 'seasons' ({seasons}) must match the number of 'season_labels' elements ({season_labels})")
+    # extract the standardized NDVI ndvi_vals within the AOI
+    ndvi_vals = std_ndvi_vals(
+        aoi_,
+        start=since_filter.isoformat(),
+        end=until_filter.isoformat()
+    )
+
+    # calculate the seasonal transition point
+    cuts = val_cuts(ndvi_vals, seasons)
+
+    # determine the seasonal time windows
+    windows = seasonal_windows(ndvi_vals, cuts, season_labels)
+
+    # Categorize the fixtime values according to the season
+    bins = pandas.IntervalIndex(data=windows.apply(lambda x: pandas.Interval(x["start"], x["end"]), axis=1))
+    labels = windows.season
+    df[index_name] = pandas.cut(df[time_col], bins=bins).map(dict(zip(bins,labels)))
+
+    # set the index
+    return df.set_index(index_name, append=True)

ecoscope/base/utils.py

@@ -212,3 +212,52 @@ def create_modis_interval_index(start, intervals, overlap=pd.Timedelta(0), close
     left = pd.DatetimeIndex(left)
 
     return pd.IntervalIndex.from_arrays(left=left, right=left + pd.Timedelta(days=16), closed=closed)
+
+
+def add_val_index(df, index_name, val):
+    """
+    Parameters
+    ----------
+    df : pd.DataFrame
+        Dataframe to create index on
+    index_name : str
+        Name of index
+    val : str
+        Current column to rename and set as index
+
+    Returns
+    -------
+    pd.DataFrame
+
+    """
+    if val in df.columns:
+        return df.rename(columns={val: index_name}).set_index(index_name, append=True)
+    else:
+        df[index_name] = val
+        df = df.set_index(index_name, append=True)
+    return df
+
+
+def add_temporal_index(df, index_name, time_col, directive):
+    """
+    Parameters
+    ----------
+    df : pd.DataFrame
+        Dataframe to create index on
+    index_name : str
+        Name of temporal index
+    time_col : str
+        Name of the time column
+    directive : str
+        Time format string
+
+    Returns
+    -------
+    pd.DataFrame
+
+    """
+    if directive and time_col:
+        df[index_name] = df[time_col].dt.strftime(directive)
+        return df.set_index(index_name, append=True)
+    else:
+        return df