#!/usr/bin/env python3

from bs4 import BeautifulSoup
from multiprocessing import Pool
import ntpath
import os
import csv
from filing import Filing

# Set directories
root_dir = os.path.dirname(os.path.abspath(__file__))
filings_dir = os.path.join(root_dir, "filings")
clean_filings_dir = os.path.join(root_dir, "cleanedfilings")
print("root_dir:", root_dir)
print("filings_dir:", filings_dir)
print("clean_filings_dir:", clean_filings_dir)


def make_batches(lst, batch_size):
    res = []
    batch = []
    for i in range(len(lst)):
        batch.append(lst[i])
        if (i + 1) % batch_size == 0:
            res.append(batch)
            batch = []

    if len(batch) > 0:
        res.append(batch)

    return res


# Read CSV data into a list of dicts
csv_filename = os.path.join(root_dir, "data", "progress.csv")
print("Reading CSV data from ", csv_filename)
csv_data_rows = []
with open(csv_filename, newline="") as csvfile:
    try:
        companyReader = csv.reader(csvfile, delimiter=";")
        for row in companyReader:
            csv_data_rows.append({
                "ticker": row[0],
                "cik": row[1],
            })
    except Exception as ex:
        print("Reading CSV failed!")
        raise ex

# Construct a list of data_objects (dicts) that contain all necessary info
# for processing.
data_object_lst = []
print("Writing filings to disk and building data objects")
for row in csv_data_rows:
    ticker = row["ticker"]
    cik = row["cik"]

    filings = Filing(cik=cik, filing_type="10-k", count=45)

    input_file_dir = os.path.join(filings_dir, cik, "10-k")

    input_filenames = []
    for r, d, f in os.walk(input_file_dir):
        for file_name in f:
            input_filenames.append(os.path.join(r, file_name))

    for input_filename in input_filenames:
        output_filename = os.path.join(
            clean_filings_dir, ticker, ntpath.basename(input_filename))

        data_object = {
            "ticker": ticker,
            "cik": cik,
            "filings_obj": filings,
            "input_filename": input_filename,
            "output_filename": output_filename,
            "input_string": None,
            "output_string": None,
        }
        data_object_lst.append(data_object)

# Multiprocessing worker function
# Stores result in data objects


def process(data_object):
    input_string = data_object["input_string"]
    output_string = BeautifulSoup(input_string, "lxml").text
    data_object["output_string"] = output_string
    return data_object


def process_data_object_batch(batch):
    # Fetch filings over HTTP
    for data_object in batch:
        filings = data_object["filings_obj"]
        filings.save(filings_dir)
        del data_object["filings_obj"]

    # Read files and store content in data objects
    print("Reading filing data into memory")
    for data_object in batch:
        input_filename = data_object["input_filename"]
        with open(input_filename, "r+") as rawFile:
            input_file_string = rawFile.read()
            data_object["input_string"] = input_file_string

    print("Parsing filing data with lxml")

    # Multiprocessing, 4 cores
    with Pool(4) as pool:
        result_batch = pool.map(process, batch)

    # No multiprocessing, does the same thing
    # result_batch = map(process, batch)

    # Write output to disk from data objects
    print("Writing filing data to disk")
    for data_object in result_batch:
        output_filename = data_object["output_filename"]
        output_dir = os.path.dirname(output_filename)
        os.makedirs(output_dir, exist_ok=True)
        with open(output_filename, "w") as newFile:
            output_string = data_object["output_string"]
            newFile.write(output_string)

    for data_object in result_batch:
        data_object.clear()  # Clean up memory


batches = make_batches(data_object_lst, 5)

for i, batch in enumerate(batches):
    print("=== Processing batch #{}".format(i))
    process_data_object_batch(batch)

print("Done.")