timothyrenner

import streamlit as st
import pandas as pd
import altair as alt
import pydeck as pdk
import os

from dateutil.parser import parse

try:
    from dotenv import load_dotenv, find_dotenv

timothyrenner

from h3 import h3
from pygfl.easy import solve_gfl

def build_neighbor_edges(hexids):
    # Hash the hexid to the position so we can easily look
    # up where the original hexid position is in the array.
    hexid_to_position = {h:ii for ii,h in enumerate(hexids)}
    edges = []
    for h in hexids:
        for n in h3.k_ring(h,1):

timothyrenner

data_frame.withColumn(
    "prediction",
    predict_pandas_udf(col("feature1"), col("feature2"), ...)
)

timothyrenner

import pandas as pd

from pyspark.sql.functions import pandas_udf
from pyspark.sql.types import DoubleType
 
 
@pandas_udf(returnType=DoubleType())
def predict_pandas_udf(*features):
    """ Executes the prediction using numpy arrays.
         

timothyrenner

my_data.rdd.mapPartitions(predict_partition).toDF()

timothyrenner

import pandas as pd
 
 
# We'll need this handy list more than once. It enforces the
# column order required by the model.
FEATURES = ["feature1", "feature2", "feature3", ...]
 
 
def predict_partition(rows):
    """ Calls a vectorized prediction by loading the partition into memory.

timothyrenner

my_df.withColumn(
    "predicted_score",
    predict_udf(col("feature1"), col("feature2"), ...)
)

timothyrenner

from pyspark.sql.functions import udf
from pyspark.sql.types import DoubleType
 
 
predict_udf = udf(predict, DoubleType())

timothyrenner

import numpy as np

def predict(*features):
    """ Performs a prediction on the features.
    
        Parameters
        ----------
        features : List[float]
            The feature values the model needs to make a prediction.
           

timothyrenner

def average_agreement(list1, list2, max_depth):
 
    # Empty lists evaluate to false.
    if (not list1) or (not list2):
        return 0.0
     
    ### NEW CODE ###
    # Truncate the depth
    max_list_len = max(len(list1), len(list2))
    max_depth = min(max_depth, max_list_len)