Cache Round-Trip Example with Metadata

This notebook demonstrates how to use the enhanced caching system that:

• Caches DataFrames based on selected function arguments

• Saves and reloads cached content via CSV with full metadata

• Supports recovery of clean data frames with datetime indexing

Features

• Metadata headers: cached_function, index_name, keys, and col_keys

• Optional preservation of some key fields in data via col_keys

• Automatic datetime index recovery on reload

import pandas as pd
from dms_datastore.caching import *

@cache_dataframe(key_args=["station", "variable"])
def ec_data(station, variable, subloc="upper"):
    # Simulated time series
    idx = pd.date_range("2020-01-01", periods=100, freq="D")
    df = pd.DataFrame({"value": range(100), "station": station, "param": variable}, index=idx)
    return df

df1 = ec_data(station="vns", variable="flow")
df2 = ec_data(station="vns", variable="ec")
df3 = ec_data(station="old", variable="ec")

cache_to_csv()  # Writes individual CSVs with function-specific metadata

LocalCache.instance().clear()
load_cache_csv("ec_data.csv")

df_reloaded = LocalCache.instance()[generate_cache_key("ec_data", station="vns", variable="flow")]
df_reloaded = coerce_datetime_index(df_reloaded)
print(df_reloaded.dtypes)
df_reloaded.head()

value         int64
station.1    object
param        object
dtype: object

	value	station.1	param
datetime
2020-01-01	0	vns	flow
2020-01-02	1	vns	flow
2020-01-03	2	vns	flow
2020-01-04	3	vns	flow
2020-01-05	4	vns	flow