Add params + better functionalities/files split

.pre-commit-config.yaml

@@ -5,7 +5,7 @@ repos:
   - id: end-of-file-fixer
   - id: trailing-whitespace
 - repo: https://github.com/psf/black
-  rev: 24.4.0
+  rev: 24.4.1
   hooks:
   - id: black
 - repo: https://github.com/PyCQA/isort

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "etrade_tax_poland"
-version = "0.0.8"
+version = "0.0.9"
 authors = [
   { name="Aleksander Kantak", email="[email protected]" },
 ]

src/etrade_tax_poland/__main__.py

@@ -1,27 +1,19 @@
 """Read all E*Trade files and parse."""
 
-import os
-import sys
-
-from .common import merge_csvs
+from .args import parse_args
 from .dividends import process_dividend_docs
+from .files_handling import merge_csvs
 from .stocks import process_stock_docs
 
 
-def parse_all_docs():
+def parse_all_docs(dir_path):
     """Figure out directory and run all functions on it."""
-    dir_path = "."
-    if len(sys.argv) > 1:
-        dir_path = sys.argv[1]
-        if not os.path.isdir(dir_path):
-            print("Provided path is not a directory")
-            sys.exit(1)
-    dir_path = os.path.abspath(dir_path)
     process_dividend_docs(dir_path)
     process_stock_docs(dir_path)
 
 
 if __name__ == "__main__":
-
-    parse_all_docs()
-    merge_csvs()
+    args = parse_args()
+    parse_all_docs(args.dirpath)
+    if not args.no_xlsx:
+        merge_csvs()

src/etrade_tax_poland/args.py

@@ -0,0 +1,18 @@
+"""Parse CLI arguments."""
+
+import argparse
+import os
+import sys
+
+
+def parse_args():
+    """Parse CLI arguments."""
+    parser = argparse.ArgumentParser()
+    parser.add_argument("dirpath", nargs="?", default=".", help="Get statements path")
+    parser.add_argument("-x", "--no-xlsx", action="store_true")
+    args = parser.parse_args()
+    if not os.path.isdir(args.dirpath):
+        print("Provided path is not a directory")
+        sys.exit(1)
+    args.dirpath = os.path.abspath(args.dirpath)
+    return args

src/etrade_tax_poland/common.py

@@ -1,132 +1,3 @@
-"""Implement common functions for etrade documents processing."""
-
-import datetime
-import glob
-import json
-import os
-from time import sleep
-
-import pyexcel.cookbook as xlsck
-import requests
-from pypdf import PdfReader
+"""Common variables and functions."""
 
 TAX_PL = 0.19
-
-
-def pdfs_in_dir(directory):
-    """Get all PDF statements files."""
-    os.chdir(directory)
-    return glob.glob("*.pdf")
-
-
-def file_to_text(filename):
-    """Parse PDF file to text only."""
-    reader = PdfReader(filename)
-    text = ""
-    for page in reader.pages:
-        text += page.extract_text() + "\n"
-    return text
-
-
-def merge_csvs():
-    """Merge csvs into xlsx and remove them."""
-    files_list = glob.glob("*.csv")
-    xlsck.merge_all_to_a_book(files_list, "etrade.xlsx")
-    for file_name in files_list:
-        os.remove(file_name)
-
-
-class NbpRatiosCache:
-    """Cache data instead of always requesting from NBP."""
-
-    date_format = "%Y-%m-%d"
-    nbp_url = "https://api.nbp.pl/api/exchangerates/rates/a/usd/{}/?format=json"
-
-    def __init__(self):
-        """Initialize objects and fields."""
-        self.cache_file = f"{os.path.dirname(os.path.abspath(__file__))}/nbp_cache.json"
-        self.read_cache()
-
-    def read_cache(self):
-        """Read cache file."""
-        if not os.path.isfile(self.cache_file):
-            self.cache = {"_": ""}
-            return
-        with open(self.cache_file, "r", encoding="utf-8") as file:
-            self.cache = json.load(file)
-
-    def write_cache(self):
-        """Write cache file."""
-        json_dump_params = {
-            "sort_keys": True,
-            "indent": 2,
-            "separators": (",", ": "),
-        }
-        with open(self.cache_file, "w", encoding="utf-8") as file:
-            json.dump(self.cache, file, **json_dump_params)
-            file.write("\n")
-
-    def get_ratio(self, date_obj):
-        """Get ratio from cache if available, otherwise request from NBP."""
-        key = date_obj.strftime(self.date_format)
-        if key in self.cache:
-            if not self.cache[key]:
-                raise ValueError("Ratio for selected date is not available in NBP")
-            return self.cache[key]
-
-        while True:
-            try:
-                current_url = self.nbp_url.format(date_obj.strftime(self.date_format))
-                req = requests.get(current_url, timeout=5)
-            except (requests.exceptions.ProxyError, requests.exceptions.ConnectTimeout):
-                print("Timeout 5s when getting USD/PLN ratio, retrying after 1 second")
-                sleep(1)
-                continue
-            if req.status_code == 200:
-                self.cache[key] = req.json()["rates"][0]["mid"]
-                self.write_cache()
-                return self.cache[key]
-            if req.status_code == 404:
-                self.cache[key] = ""
-                self.write_cache()
-                raise ValueError("Ratio for selected date is not available in NBP")
-            print(f"{req.status_code} {req.text}")
-            print("Unhandled error when getting USD/PLN ratio, retrying after 1 second")
-            sleep(1)
-            continue
-
-
-NBP_CACHE = NbpRatiosCache()
-
-
-def date_to_usd_pln(date_obj):
-    """Find 'day before vestment' USD/PLN ratio."""
-    while True:
-        date_obj -= datetime.timedelta(days=1)
-        try:
-            ratio = NBP_CACHE.get_ratio(date_obj)
-            break
-        except ValueError:
-            date_obj -= datetime.timedelta(days=1)
-    return (date_obj, ratio)
-
-
-def save_csv(filename, header, lines):
-    """Save header and lines to a csv file."""
-    if not lines:
-        return
-    with open(filename, "w", encoding="utf-8") as file:
-        file.write(f"{header}\n")
-        for line in lines:
-            file.write(f"{line}\n")
-
-
-def sum_header():
-    """Return sum csv file header."""
-    return ",".join(
-        [
-            "NAME",
-            "VALUE",
-            "PIT_FIELD",
-        ]
-    )

src/etrade_tax_poland/dividends.py

@@ -2,7 +2,9 @@
 
 import datetime
 
-from . import common as etc
+from . import files_handling as fh
+from . import nbp
+from .common import TAX_PL
 
 
 class Dividend:
@@ -64,7 +66,7 @@ def insert_currencies_ratio(self, ratio_date, ratio_value):
         self.ratio_date = ratio_date
         self.ratio_value = ratio_value
         self.pln_gross = round(self.usd_gross * ratio_value, 2)
-        self.flat_rate_tax = round(self.pln_gross * etc.TAX_PL, 2)
+        self.flat_rate_tax = round(self.pln_gross * TAX_PL, 2)
         self.pln_tax_paid = round(self.usd_tax * ratio_value, 2)
         self.pln_tax_due = self.flat_rate_tax - self.pln_tax_paid
 
@@ -139,22 +141,22 @@ def divs_sum_csved(dividends):
 
 def process_dividend_docs(directory):
     """Count due tax based on statements files in directory."""
-    files = etc.pdfs_in_dir(directory)
+    files = fh.pdfs_in_dir(directory)
     dividends = []
     for filename in files:
-        text = etc.file_to_text(f"{directory}/{filename}")
+        text = fh.file_to_text(f"{directory}/{filename}")
         if dividend := get_stock_dividend_from_text(text):
             dividend.file = filename
-            dividend.insert_currencies_ratio(*etc.date_to_usd_pln(dividend.pay_date))
+            dividend.insert_currencies_ratio(*nbp.date_to_usd_pln(dividend.pay_date))
             dividends.append(dividend)
         if ldivs := get_liquidity_dividends_from_text(text):
             for ldiv in ldivs:
                 ldiv.file = filename
-                ldiv.insert_currencies_ratio(*etc.date_to_usd_pln(ldiv.pay_date))
+                ldiv.insert_currencies_ratio(*nbp.date_to_usd_pln(ldiv.pay_date))
             dividends += ldivs
 
     for dividend in dividends:
-        dividend.insert_currencies_ratio(*etc.date_to_usd_pln(dividend.pay_date))
+        dividend.insert_currencies_ratio(*nbp.date_to_usd_pln(dividend.pay_date))
 
-    etc.save_csv("_dividend.csv", Dividend.csv_header(), [d.csved() for d in dividends])
-    etc.save_csv("dividends_summary.csv", etc.sum_header(), divs_sum_csved(dividends))
+    fh.save_csv("_dividend.csv", Dividend.csv_header(), [d.csved() for d in dividends])
+    fh.save_csv("dividends_summary.csv", fh.sum_header(), divs_sum_csved(dividends))

src/etrade_tax_poland/files_handling.py

@@ -0,0 +1,51 @@
+"""Implement common functions for files processing."""
+
+import glob
+import os
+
+import pyexcel.cookbook as xlsck
+from pypdf import PdfReader
+
+
+def pdfs_in_dir(directory):
+    """Get all PDF statements files."""
+    os.chdir(directory)
+    return glob.glob("*.pdf")
+
+
+def file_to_text(filename):
+    """Parse PDF file to text only."""
+    reader = PdfReader(filename)
+    text = ""
+    for page in reader.pages:
+        text += page.extract_text() + "\n"
+    return text
+
+
+def save_csv(filename, header, lines):
+    """Save header and lines to a csv file."""
+    if not lines:
+        return
+    with open(filename, "w", encoding="utf-8") as file:
+        file.write(f"{header}\n")
+        for line in lines:
+            file.write(f"{line}\n")
+
+
+def merge_csvs():
+    """Merge csvs into xlsx and remove them."""
+    files_list = glob.glob("*.csv")
+    xlsck.merge_all_to_a_book(files_list, "etrade.xlsx")
+    for file_name in files_list:
+        os.remove(file_name)
+
+
+def sum_header():
+    """Return sum csv file header."""
+    return ",".join(
+        [
+            "NAME",
+            "VALUE",
+            "PIT_FIELD",
+        ]
+    )

src/etrade_tax_poland/nbp.py

@@ -0,0 +1,83 @@
+"""Implement NBP currencies ratios gathering."""
+
+import datetime
+import json
+import os
+from time import sleep
+
+import requests
+
+
+class NbpRatiosCache:
+    """Cache data instead of always requesting from NBP."""
+
+    date_format = "%Y-%m-%d"
+    nbp_url = "https://api.nbp.pl/api/exchangerates/rates/a/usd/{}/?format=json"
+
+    def __init__(self):
+        """Initialize objects and fields."""
+        self.cache_file = f"{os.path.dirname(os.path.abspath(__file__))}/nbp_cache.json"
+        self.read_cache()
+
+    def read_cache(self):
+        """Read cache file."""
+        if not os.path.isfile(self.cache_file):
+            self.cache = {"_": ""}
+            return
+        with open(self.cache_file, "r", encoding="utf-8") as file:
+            self.cache = json.load(file)
+
+    def write_cache(self):
+        """Write cache file."""
+        json_dump_params = {
+            "sort_keys": True,
+            "indent": 2,
+            "separators": (",", ": "),
+        }
+        with open(self.cache_file, "w", encoding="utf-8") as file:
+            json.dump(self.cache, file, **json_dump_params)
+            file.write("\n")
+
+    def get_ratio(self, date_obj):
+        """Get ratio from cache if available, otherwise request from NBP."""
+        key = date_obj.strftime(self.date_format)
+        if key in self.cache:
+            if not self.cache[key]:
+                raise ValueError("Ratio for selected date is not available in NBP")
+            return self.cache[key]
+
+        while True:
+            try:
+                current_url = self.nbp_url.format(date_obj.strftime(self.date_format))
+                req = requests.get(current_url, timeout=5)
+            except (requests.exceptions.ProxyError, requests.exceptions.ConnectTimeout):
+                print("Timeout 5s when getting USD/PLN ratio, retrying after 1 second")
+                sleep(1)
+                continue
+            if req.status_code == 200:
+                self.cache[key] = req.json()["rates"][0]["mid"]
+                self.write_cache()
+                return self.cache[key]
+            if req.status_code == 404:
+                self.cache[key] = ""
+                self.write_cache()
+                raise ValueError("Ratio for selected date is not available in NBP")
+            print(f"{req.status_code} {req.text}")
+            print("Unhandled error when getting USD/PLN ratio, retrying after 1 second")
+            sleep(1)
+            continue
+
+
+NBP_CACHE = NbpRatiosCache()
+
+
+def date_to_usd_pln(date_obj):
+    """Find 'day before vestment' USD/PLN ratio."""
+    while True:
+        date_obj -= datetime.timedelta(days=1)
+        try:
+            ratio = NBP_CACHE.get_ratio(date_obj)
+            break
+        except ValueError:
+            date_obj -= datetime.timedelta(days=1)
+    return (date_obj, ratio)