Implement fuzzy category input

src/fzf.py

@@ -0,0 +1,126 @@
+from __future__ import print_function
+
+import errno
+import os.path
+import subprocess
+import sys
+
+from pkg_resources import resource_exists, resource_filename
+
+__all__ = 'BUNDLED_EXECUTABLE', 'iterfzf'
+
+EXECUTABLE_NAME = 'fzf.exe' if sys.platform == 'win32' else 'fzf'
+BUNDLED_EXECUTABLE = (
+    resource_filename(__name__, EXECUTABLE_NAME)
+    if resource_exists(__name__, EXECUTABLE_NAME)
+    else (
+        os.path.join(os.path.dirname(__file__), EXECUTABLE_NAME)
+        if os.path.isfile(
+            os.path.join(os.path.dirname(__file__), EXECUTABLE_NAME)
+        )
+        else None
+    )
+)
+
+
+def iterfzf(
+    # CHECK: When the signature changes, __init__.pyi file should also change.
+    iterable,
+    # Search mode:
+    extended=True, exact=False, case_sensitive=None,
+    # Interface:
+    multi=False, mouse=True, print_query=False,
+    # Layout:
+    prompt='> ',
+    ansi=None,
+    preview=None,
+    # Misc:
+    query='', encoding=None, executable=BUNDLED_EXECUTABLE or EXECUTABLE_NAME
+):
+    cmd = [executable, '--no-sort', '--prompt=' + prompt]
+    cmd = [executable, '--prompt=' + prompt]
+    if not extended:
+        cmd.append('--no-extended')
+    if case_sensitive is not None:
+        cmd.append('+i' if case_sensitive else '-i')
+    if exact:
+        cmd.append('--exact')
+    if multi:
+        cmd.append('--multi')
+    if not mouse:
+        cmd.append('--no-mouse')
+    if print_query:
+        cmd.append('--print-query')
+    if query:
+        cmd.append('--query=' + query)
+    if preview:
+        cmd.append('--preview=' + preview)
+    if ansi:
+        cmd.append('--ansi')
+    encoding = encoding or sys.getdefaultencoding()
+    proc = None
+    stdin = None
+    byte = None
+    lf = u'\n'
+    cr = u'\r'
+    for line in iterable:
+        if byte is None:
+            byte = isinstance(line, bytes)
+            if byte:
+                lf = b'\n'
+                cr = b'\r'
+        elif isinstance(line, bytes) is not byte:
+            raise ValueError(
+                'element values must be all byte strings or all '
+                'unicode strings, not mixed of them: ' + repr(line)
+            )
+        if lf in line or cr in line:
+            raise ValueError(r"element values must not contain CR({1!r})/"
+                             r"LF({2!r}): {0!r}".format(line, cr, lf))
+        if proc is None:
+            proc = subprocess.Popen(
+                cmd,
+                stdin=subprocess.PIPE,
+                stdout=subprocess.PIPE,
+                stderr=None
+            )
+            stdin = proc.stdin
+        if not byte:
+            line = line.encode(encoding)
+        try:
+            stdin.write(line + b'\n')
+            stdin.flush()
+        except IOError as e:
+            if e.errno != errno.EPIPE and errno.EPIPE != 32:
+                raise
+            break
+    stdin.close()
+    if proc is None or proc.wait() not in [0, 1]:
+        if print_query:
+            return None, None
+        else:
+            return None
+    try:
+        stdin.close()
+    except IOError as e:
+        if e.errno != errno.EPIPE and errno.EPIPE != 32:
+            raise
+    stdout = proc.stdout
+    decode = (lambda b: b) if byte else (lambda t: t.decode(encoding))
+    output = [decode(ln.strip(b'\r\n\0')) for ln in iter(stdout.readline, b'')]
+    if print_query:
+        try:
+            if multi:
+                return output[0], output[1:]
+            else:
+                return output[0], output[1]
+        except IndexError:
+            return output[0], None
+    else:
+        if multi:
+            return output
+        else:
+            try:
+                return output[0]
+            except IndexError:
+                return None

src/models.py

@@ -4,6 +4,9 @@ from pathlib import Path
 from typing import List
 
 
+UNKNOWN_CATEGORY = 'account2'
+
+
 class CsvConfig(BaseModel):
     """
     Class to define how to parse a certain CSV file. We use the

src/predict.py

@@ -0,0 +1,26 @@
+from src.models import Transaction, UNKNOWN_CATEGORY
+from src.fzf import iterfzf
+from typing import List
+
+
+def get_categories(transactions: List[Transaction]) -> List[str]:
+    categories = set([t.account2 for t in transactions])
+    categories.discard(UNKNOWN_CATEGORY)
+    return list(categories)
+
+
+def add_account2(transactions: List[Transaction]):
+    categories = get_categories(transactions)
+    unmapped_transactions = filter(lambda t: t.account2 == UNKNOWN_CATEGORY, transactions)
+    for t in unmapped_transactions:
+        add_account2_interactive(t, categories)
+
+
+def add_account2_interactive(transaction: Transaction, categories: List[str]):
+    t = transaction
+    account2 = None
+    prompt = f"{t.account1} {t.date} {t.description} {t.debit} > "
+    while account2 is None:
+        account2 = iterfzf(categories, prompt=prompt)
+    transaction.account2 = account2
+    print(f"Assigned category '{account2}'.")

src/process.py

@@ -5,6 +5,8 @@ import sys
 import datetime
 import src.utils
 import src.write
+import src.models
+import src.predict
 from src.models import Config, CsvConfig, Transaction
 from typing import List, Dict
 
@@ -47,7 +49,7 @@ def get_transactions(csv_file: str, config: CsvConfig) -> List[Transaction]:
                 credit=amount,
                 date=date_to_date(t['date']),
                 account1=config.account1,
-                account2="account2",
+                account2=src.models.UNKNOWN_CATEGORY,
                 description=t['description'],
                 csv_file=csv_file,
                 row=csv_file + ", " + ", ".join(row))
@@ -97,6 +99,7 @@ def process_csv_files(config: Config):
     find_duplicates(transactions)
     mappings = src.utils.read_mappings(config.mappings_file)
     apply_mappings(transactions, mappings)
+    src.predict.add_account2(transactions)
     src.utils.write_mappings(transactions, config.mappings_file)
     src.write.render_to_file(transactions, config.output_file)