Module 1, Practical 10¶

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
import argparse


def get_longest_call(calls):
    longest_call = calls.loc[calls['duration'].idxmax()]
    id_caller, id_callee = longest_call.caller, longest_call.callee
    return id_caller, id_callee


def get_friends_fb(fb):
    fb_friends = []
    for _, i in fb.iterrows():
        user_a = i[0]
        user_b = i[1]
        fb_friends.append((user_a, user_b))
    return fb_friends


def get_gender(gender, id_caller, id_callee):
    print("Gender of", id_callee, "=", gender.get(id_callee, "?"))
    print("Gender of", id_caller, "=", gender.get(id_caller, "?"))


def get_sms_count_dict(sms):
    sms_count = dict()
    for _, i in sms.iterrows():
        _, sender, receiver = i
        if sender > receiver:
            sender, receiver = receiver, sender
        sms_count[(sender, receiver)] = sms_count.get((sender, receiver), 0) + 1
    return sms_count


def get_tot_messages(sms_count, fb_friends):
    tot_messages = 0
    for k, v in sms_count.items():
        id_a, id_b = k
        if (id_a, id_b) not in fb_friends and (id_b, id_a) not in fb_friends:
            tot_messages += v
    return tot_messages


def load_gender(path):
    with open(path, "r") as f:
        data = f.read()
    res = {}
    for el in data.split(" "):
        tmp = el.split("_")
        if len(tmp) == 2:
            idx, gender = tmp
            res[int(idx)] = gender.strip()
    return res


def get_discretized_times(calls_small):
    times_call = []
    for t, a, b, d in calls_small.to_numpy():
        times_call.append(int(t / (60 * 60)))
    return times_call


def my_plot(times_call, plot_name="nb_interactions.png"):
    t_c, count_c = np.unique(times_call, return_counts=True)
    x_ticks = []
    x_labels = []
    for i in range(int((max(t_c) / 24) + 1)):
        x_ticks.append(i * 24)
        x_labels.append("Day " + str(i + 1))
    plt.figure(figsize=(15, 4))
    plt.plot(t_c, count_c, label="calls", color="blue")
    plt.ylabel("Number of interactions")
    plt.xticks(x_ticks, x_labels, rotation=45)
    plt.title("Calls Over Time")
    plt.legend()
    plt.tight_layout()
    plt.savefig(plot_name)
    print(f"Plot saved as '{plot_name}'")
    plt.show()


def Top3_Persistent_Callers(calls, gender_dict):
    calls = calls.sort_values(by="timestamp")
    pair_timestamps = {}
    for _, row in calls.iterrows():
        a, b, t = row["caller"], row["callee"], row["timestamp"]
        pair = tuple(sorted((a, b)))
        pair_timestamps.setdefault(pair, []).append(t)

    persistent_per_pair = {}
    for pair, times in pair_timestamps.items():
        times.sort()
        count = sum(abs(times[i + 1] - times[i]) <= 7200 for i in range(len(times) - 1))
        persistent_per_pair[pair] = count

    persistent_score = {}
    for (a, b), score in persistent_per_pair.items():
        persistent_score[a] = persistent_score.get(a, 0) + score
        persistent_score[b] = persistent_score.get(b, 0) + score

    top3 = sorted(persistent_score.items(), key=lambda x: x[1], reverse=True)[:3]
    print("\nTop 3 Persistent Callers:")
    print("--------------------------")
    for sid, score in top3:
        print(f"Student {sid}:")
        print(f"   Persistent Score: {score}")
        print(f"   Gender: {gender_dict.get(sid, '?')}\n")
    return top3


# ======================
# ARGPARSE ENTRY POINT
# ======================
def main():
    parser = argparse.ArgumentParser(description="Analyze DTU communication dataset.")

    parser.add_argument(
        "--calls", default="DTU/calls.csv", help="Path to calls.csv file (default: DTU/calls.csv)"
    )
    parser.add_argument(
        "--fb", default="DTU/fb_friends.csv", help="Path to fb_friends.csv file (default: DTU/fb_friends.csv)"
    )
    parser.add_argument(
        "--genders", default="DTU/genders.txt", help="Path to genders.txt file (default: DTU/genders.txt)"
    )
    parser.add_argument(
        "--plot_name",
        default="nb_interactions.png",
        help="Filename for saving the call plot (default: nb_interactions.png)",
    )

    args = parser.parse_args()

    # Load data
    fb = pd.read_csv(args.fb)
    calls = pd.read_csv(args.calls)
    gender = load_gender(args.genders)

    # 1) Longest call
    id_caller, id_callee = get_longest_call(calls)
    print("1) The longest call is between:", id_caller, "and", id_callee)

    # 2) Check if they are FB friends
    fb_friends = get_friends_fb(fb)
    if (id_callee, id_caller) in fb_friends or (id_caller, id_callee) in fb_friends:
        print("2)", id_callee, "and", id_caller, "are friends on FB")

    # 3) Gender
    print("3)")
    get_gender(gender, id_caller, id_callee)

    # 4) Plot calls over time
    print("4) Plotting call activity over time...")
    times_call = get_discretized_times(calls)
    my_plot(times_call, plot_name=args.plot_name)

    # 5) Persistent Score
    print("5) Persistent Score analysis")
    Top3_Persistent_Callers(calls, gender)


if __name__ == "__main__":
    main()