add split_datetime (#1811)

add split_datetime

Made in [Fused Workbench](https://www.fused.io/workbench)

---------

Co-authored-by: sina@fused.io <sina@fused.io>
Co-authored-by: Sina Kashuk <sina@nyu.edu>

Author: fused-io-unstable[bot]

public/common/common.py

@@ -2440,6 +2440,64 @@ def from_path(path: str):
 
     return obj
 
+
+def parse_date(date_str: str, is_end: bool = False):
+    """
+    Parse a partial or full date string and return ISO8601 with UTC 'Z'.
+    Handles year-only, year-month, and full dates.
+    """
+    from datetime import timedelta
+    import pandas as pd
+
+    # Handle partial dates
+    if len(date_str) == 4:  # Year only
+        date_str = f"{date_str}-12-31" if is_end else f"{date_str}-01-01"
+    elif len(date_str) == 7:  # Year-month
+        y, m = date_str.split('-')
+        if is_end:
+            last_day = pd.Timestamp(f"{y}-{m}-01").days_in_month
+            date_str = f"{y}-{m}-{last_day:02d}"
+        else:
+            date_str = f"{y}-{m}-01"
+
+    # Parse with pandas (handles most formats automatically)
+    dt = pd.to_datetime(date_str, errors='coerce', utc=True)
+    if pd.isna(dt):
+        raise ValueError(f"Could not parse date: {date_str}")
+
+    # Set end_date-of-day if needed
+    if is_end and dt.hour == 0 and dt.minute == 0 and dt.second == 0:
+        dt = dt.replace(hour=23, minute=59, second=59)
+
+    return dt
+
+
+def split_datetime(start_date: str, end_date: str, n: int = 10, gap_milliseconds:int = 1000):
+    """
+    Split start_date/end_date dates into n chunks, returning list of (start_iso, end_iso) strings.
+    """
+    from datetime import timedelta
+    start_dt = parse_date(start_date, is_end=False)
+    end_dt = parse_date(end_date, is_end=True)
+
+    total_seconds = (end_dt - start_dt).total_seconds()
+    chunk_seconds = total_seconds / n
+
+    chunks = []
+    for i in range(n):
+        chunk_start = start_dt + timedelta(seconds=i * chunk_seconds)
+        chunk_end = start_dt + timedelta(seconds=(i + 1) * chunk_seconds)
+        if i < n - 1:
+            chunk_end = chunk_end - timedelta(milliseconds=gap_milliseconds)
+        
+        chunks.append((
+            chunk_start.strftime("%Y-%m-%dT%H:%M:%SZ"),
+            chunk_end.strftime("%Y-%m-%dT%H:%M:%SZ")
+        ))
+
+    return chunks
+
+
 def split_gdf(gdf, n: int = None, zoom: int = None, clip: bool = False, return_type: str = "gdf"):
     import geopandas as gpd
     import pickle
@@ -3669,7 +3727,7 @@ def estimate_zoom(bounds, target_num_tiles=1):
 
 
 def get_tiles(
-    bounds=None, target_num_tiles=1, zoom=None, max_tile_recursion=6, as_gdf=True, clip=False, verbose=False
+    bounds=None, target_num_tiles=1, zoom=None, max_tile_recursion=6, as_gdf=True, clip=False, add_bounds=False, verbose=False
 ):
     bounds = to_gdf(bounds)
     import mercantile
@@ -3724,6 +3782,8 @@ def get_tiles(
         if verbose: 
             print(f"Generated {len(gdf)} tiles at zoom level {zoom_level}")
 
+    if add_bounds:
+        gdf["bounds"] = gdf["geometry"].map(lambda x: list(x.bounds))
     if clip:
         return gdf.clip(bounds) if as_gdf else gdf[["x", "y", "z"]].values
     else: