Messenger_1Million_Stats/million/analyze/find_holes.py

from typing import List
from million.model.message import Message
from million.model.sequence import Sequence
import million.analyze.message_evaluation as msg_val


def compute_sequences(messages: List[Message], accepted_max: int = 1_000_000) -> List[Sequence]:
    """ 
    Takes a list of messages as input and returns a list of sequences
    for every following messages with following 'counted values'
    """
    sequences: List[Sequence] = [Sequence(start_message=messages[0])]
    
    for message in messages[1:]:
        if msg_val.get(message) > accepted_max: continue

        if msg_val.get(message) == sequences[-1].end() + 1:
            sequences[-1].end_message = message
        else:
            sequences.append(Sequence(start_message=message))

    return sequences            


def merge_duplicates(sequences: List[Sequence]) -> List[Sequence]:
    """ 
    Take sequences as an input and returns a list with every
    overlapping input sequences merged in one
    """
    o_sequences = sorted(sequences, key= lambda s : s.start())
    current = o_sequences[0]

    result = []
        
    for sequence in o_sequences[1:]:
        if current.overlaps(sequence):
            current.merge(sequence)
        else:
            result.append(current)
            current = sequence

    return result



def invert_sequences(sequences: List[Sequence]) -> List[Sequence]:
    """ 
    Returns the sequences representing the spaces between
    the ones given as input
    """
    result = []

    for previous, current in zip(sequences[:-1],sequences[1:]):
        result.append(Sequence(
            start_message=previous.end_message,
            end_message=current.start_message
        ))

    return result

def find_holes(messages: List[Message], accepted_max: int = 1_000_000) -> List[Sequence]:
    """
    Find the holes in the conversation
    TODO might need to be moved inside scripts/find_holes
    """
    sequences = compute_sequences(messages, accepted_max)
    merged = merge_duplicates(sequences)
    merged = [s for s in merged if s.length() > 1]

    return invert_sequences(merged)