DataFrame.append was removed in pandas 2.0

[charlesdtdb]

here is the code to correct it:

#@markdown ---
#@markdown ### Enter the trix url for the sheet file containing the Client Sales Data:
#@markdown The spreadsheet should contain the mandatory columns:
#@markdown * date: date in the format YYYY-MM-DD
#@markdown * geo: the number which identifies the geo
#@markdown * response: variable on which you want to measure incrementality
#@markdown (e.g. sales, transactions)
#@markdown * cost: variable used as spend proxy (e.g. ad spend)

#@markdown Other columns can be present in the spreadsheet.

#@markdown Spreadsheet URL containing the geo level response and spend data
client_sales_table = "https://docs.google.com/spreadsheets/d/1bXZEUasXxEqE-lE0HrBInedrdOt2zoS3TyVbkE-xQco/edit?gid=0#gid=0" #@param {type:"string"}

#@markdown Leave the following field empty if you don't want to add constraint to the geo_eligibility
geo_eligibility_table = "https://docs.google.com/spreadsheets/d/1A_J8XCqXPd4i6WJruZ_6WDHMwe7wxiILwAKhY-_GV1w/edit?gid=0#gid=0" #@param {type:"string"}
auth.authenticate_user()
creds, _ = google_auth.default()
gc = gspread.authorize(creds)
wks = gc.open_by_url(client_sales_table).sheet1
data = wks.get_all_values()
headers = data.pop(0)
geo_level_time_series = pd.DataFrame(data, columns=headers)

geo_level_time_series["date"] = pd.to_datetime(geo_level_time_series["date"])
for colname in ["response", "geo", "cost"]:
geo_level_time_series[colname] = pd.to_numeric(geo_level_time_series[colname])

num_geos = geo_level_time_series["geo"].nunique()

if not geo_eligibility_table:
geo_eligibility = None
else:
wks = gc.open_by_url(geo_eligibility_table).sheet1
data = wks.get_all_values()
headers = data.pop(0)
geo_eligibility = pd.DataFrame(data, columns=headers)
for colname in ["geo", "control", "treatment", "exclude"]:
geo_eligibility[colname] = pd.to_numeric(geo_eligibility[colname])

## build defaults for ALL geos (eligible for either if not in the sheet)
all_geos = (geo_level_time_series[["geo"]]
            .drop_duplicates()
            .assign(control=1, treatment=1, exclude=0))

geo_eligibility = all_geos.merge(
    geo_eligibility[["geo","control","treatment","exclude"]],
    on="geo", how="left", suffixes=("", "_set")
)
for col in ["control", "treatment", "exclude"]:
    geo_eligibility[col] = (
        geo_eligibility[f"{col}_set"]
        .fillna(geo_eligibility[col])  # keep defaults (1/1/0) if blank
        .fillna(0)
        .astype(int)
)
geo_eligibility.drop(columns=[c for c in geo_eligibility.columns if c.endswith("_set")],
                    inplace=True)

geo_eligibility = geoeligibility.GeoEligibility(geo_eligibility)
geo_eligibility.data.index = (
    pd.to_numeric(geo_eligibility.data.index, downcast="integer").astype(str)

)