analysis.py

import config
import matplotlib.pyplot as plt
import datetime as dt
import os.path
from instagram_interface import InstagramInterface

def get_post_info():
    post_info = {}
    post_id_lookup_by_datetime = {}
    post_count = -1

    if os.path.exists(config.account_info_file):
        post_count = InstagramInterface.load_header_info(config.account_info_file)['posts']

    with open(config.post_info_file, 'r') as fin:
        first = True
        for line in fin:
            if first:
                if line == 'link, datetime_string, likes, views, case\n':
                    first = False
                    continue
                else:
                    raise Exception('header string "link, datetime_string, likes, views, case\\n" not found.')

            line_info = {}

            line = line.replace('\n', '')
            link, datetime_string, likes, views, case = line.split(', ')

            datetime_string = datetime_string.replace('.000Z', '')
            datetime = dt.datetime.fromisoformat(datetime_string)
            likes = int(likes)
            views = int(views)
            post_id = post_count
            post_count -= 1

            line_info = {
                'link' : link,
                'datetime': datetime,
                'likes' : likes,
                'views' : views,
                'case' : case
            }

            post_info[post_id] = line_info
            post_id_lookup_by_datetime[datetime] = post_id

    return post_info, post_id_lookup_by_datetime


def compress(dates, values):
    """
    inputs:
        dates      | values
        -----------+--------
        25-12-2019 | 5
        25-12-2019 | 6
        25-12-2019 | 7
        25-12-2019 | 8
        25-12-2019 | 9
        25-12-2019 | 10
        29-12-2019 | 11
        29-12-2019 | 12
        29-12-2019 | 13
        03-01-2020 | 14
        03-01-2020 | 15
        03-01-2020 | 16
        03-01-2020 | 17

    returns:
        dates      | values
        -----------+--------
        25-12-2019 | 10
        29-12-2019 | 13
        03-01-2020 | 17
    """
    if len(dates) != len(values):
        raise Exception("Length of dates and values must match.")

    dates.sort()
    values.sort()

    compressed_dates = []
    compressed_values = []

    compresser = {}

    for d, v in zip(dates, values):
        compresser[d] = v
    
    for d, v in compresser.items():
        compressed_dates.append(d)
        compressed_values.append(v)

    return compressed_dates, compressed_values


def interpolate(dates, values):
    """
    inputs:
        dates      | values
        -----------+--------
        25-12-2019 | 5
        29-12-2019 | 8
        03-01-2020 | 10

    returns:
        dates      | values
        -----------+--------
        25-12-2019 | 5
        26-12-2019 | 5
        27-12-2019 | 5
        28-12-2019 | 5
        29-12-2019 | 8
        30-12-2019 | 8
        31-12-2019 | 8
        01-01-2020 | 8
        02-01-2020 | 8
        03-01-2020 | 10
    """
    if len(dates) != len(values):
        raise Exception("Length of dates and values must match.")

    index = 0

    dates.sort()
    values.sort()

    filled_in_dates = []
    filled_in_values = []

    while True:
        current_date = dates[index]
        current_value = values[index]
        
        filled_in_dates.append(current_date)
        filled_in_values.append(current_value)
        
        prev_value = current_value
        
        expected_next_date = current_date + dt.timedelta(days=1)
        index += 1

        if index == len(dates):
            break

        while expected_next_date != dates[index]:
            if expected_next_date > dates[index]:
                raise Exception('dates and values must be compressed before interpolation.')

            filled_in_dates.append(expected_next_date)
            filled_in_values.append(prev_value)
            expected_next_date = expected_next_date + dt.timedelta(days=1)
        
    return filled_in_dates, filled_in_values


def get_postrate(dates, values, rate):
    """
    rate: number of days

    input:
        rate: 2
        -----------+--------
        dates      | values
        -----------+--------
        25-12-2019 | 5
        26-12-2019 | 5
        27-12-2019 | 5
        28-12-2019 | 5
        29-12-2019 | 8
        30-12-2019 | 8
        31-12-2019 | 8
        01-01-2020 | 8
        02-01-2020 | 8
        03-01-2020 | 10

    returns:
        dates      | values
        -----------+--------
        25-12-2019 | 0
        26-12-2019 | 0
        27-12-2019 | 3
        28-12-2019 | 3
        29-12-2019 | 0
        30-12-2019 | 0
        31-12-2019 | 0
        01-01-2020 | 2
    """
    if len(dates) != len(values):
        raise Exception("Length of dates and values must match.")

    index = 0
    
    rate_dates = []
    rate_values = []

    while True:
        # first_date = dates[index]
        first_value = values[index]

        if index + rate >= len(dates):
            break

        next_date = dates[index + rate]
        next_value = values[index + rate]
        rate_value = next_value - first_value

        rate_dates.append(next_date)
        rate_values.append(rate_value)

        index += 1

    return rate_dates, rate_values


def normalise_values(values):
    max_v = max(values)
    min_v = min(values)
    m = 1 / (max_v - min_v)
    c = -min_v / (max_v - min_v)
    return [m * v + c for v in values]

"""
script
"""
post_info, post_id_lookup_by_datetime = get_post_info()


# dates = raw_dates
# ci_dates, postids = compress(dates, postids)
# ci_dates, postids = interpolate(ci_dates, postids)
# rdates, rates = get_postrate(ci_dates, postids, 365)
# n_postids = normalise_values(postids)
# n_rates = normalise_values(rates)
# n_likes = normalise_values(likes)

# plt.plot(ci_dates, n_postids)
# plt.plot(rdates, n_rates)
# plt.plot(raw_dates, n_likes, 'go')
# plt.show()