From 2c8b068fe2e5241a4c96d2284eddebeb4b526c05 Mon Sep 17 00:00:00 2001
From: Eddy Pedroni <epedroni@pm.me>
Date: Sun, 22 Sep 2024 18:12:06 +0200
Subject: Add new parser with tests, venv

---
 .gitignore             |  2 ++
 bootstrap-venv.sh      |  4 +++
 requirements.txt       |  2 ++
 src/parser.py          | 95 ++++++++++++++++++++++++++++++++++++++++++++++++++
 src/parser_unittest.py | 79 +++++++++++++++++++++++++++++++++++++++++
 5 files changed, 182 insertions(+)
 create mode 100644 .gitignore
 create mode 100755 bootstrap-venv.sh
 create mode 100644 requirements.txt
 create mode 100644 src/parser.py
 create mode 100644 src/parser_unittest.py

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..92afa22
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,2 @@
+__pycache__/
+venv/
diff --git a/bootstrap-venv.sh b/bootstrap-venv.sh
new file mode 100755
index 0000000..b105996
--- /dev/null
+++ b/bootstrap-venv.sh
@@ -0,0 +1,4 @@
+#!/usr/bin/zsh
+
+python -m venv venv
+./venv/bin/pip install -r requirements.txt
diff --git a/requirements.txt b/requirements.txt
new file mode 100644
index 0000000..ad90bc2
--- /dev/null
+++ b/requirements.txt
@@ -0,0 +1,2 @@
+click
+pytest
diff --git a/src/parser.py b/src/parser.py
new file mode 100644
index 0000000..db05089
--- /dev/null
+++ b/src/parser.py
@@ -0,0 +1,95 @@
+from pathlib import Path
+from collections import namedtuple
+from enum import Enum
+from typing import TextIO, Iterator
+
+Card = namedtuple('Card', ['id', 'front', 'back'])
+
+def _getCard(front_lines: list[str], back_lines: list[str]) -> Card:
+    front_text = "".join(front_lines).strip()
+    back_text = "".join(back_lines).strip()
+    id = hash(front_text + back_text)
+    return Card(id, front_text, back_text)
+
+def _getCards(f: TextIO) -> Iterator[Card]:
+    class State(Enum):
+        PARSE_FRONT = 1,
+        PARSE_BACK = 2
+    
+    state = None
+    front_lines = []
+    back_lines = []
+
+    for i, line in enumerate(f):
+        match line.strip():
+            case "FRONT":
+                # Edge case: FRONT twice in a row
+                if state == State.PARSE_FRONT:
+                    raise Exception(f"Unexpected 'FRONT': {f}:{i}")
+
+                # Next card is starting, wrap up current one
+                if state == State.PARSE_BACK:
+                    yield _getCard(front_lines, back_lines)
+                    front_lines.clear()
+                    back_lines.clear()
+
+                state = State.PARSE_FRONT
+
+            case "BACK":
+                # Edge case: BACK without FRONT before it
+                if state != State.PARSE_FRONT:
+                    raise Exception(f"Unexpected 'BACK': {f}:{i}")
+
+                state = State.PARSE_BACK
+                
+            case _:
+                match state:
+                    case State.PARSE_FRONT:
+                        front_lines += line
+                    case State.PARSE_BACK:
+                        back_lines += line
+                    # Edge case: file does not start with FRONT, flush preamble
+                    case _:
+                        continue
+
+    # Edge case: file did not end with contents of BACK
+    if state == State.PARSE_FRONT:
+        raise Exception(f"Unexpected end of file")
+
+    # Edge case: file was empty
+    if state is None:
+        return
+
+    yield _getCard(front_lines, back_lines)
+
+def parse(path: Path) -> list[Card]:
+    """
+    Parse a .fcard file and return a list of Card instances.
+
+    The parser expects .fcard files in the following format:
+
+    FRONT
+    This is the front of the first card.
+
+    BACK
+    This is the back of the first card.
+
+    FRONT
+    This is another card.
+
+    Multiple lines on the front are allowed.
+
+    BACK
+    Multiple lines on the back?
+
+    Also allowed.
+
+    FRONT
+    ...
+    """
+    if not path.is_file():
+        print(f"[Warning] Not a file: {path}")
+        return []
+
+    with open(path, "r") as f:
+        return [card for card in _getCards(f)]
diff --git a/src/parser_unittest.py b/src/parser_unittest.py
new file mode 100644
index 0000000..02638e6
--- /dev/null
+++ b/src/parser_unittest.py
@@ -0,0 +1,79 @@
+import pytest
+import parser
+from pathlib import Path
+
+# Happy path
+def test_validFile(tmp_path):
+    file_contents = """
+
+FRONT
+
+
+Foo
+
+Bar
+
+BACK
+
+
+Fizz
+
+Buzz
+
+
+
+FRONT
+
+Another card
+
+BACK
+
+Another back
+
+
+        """
+    expected = [
+        ("Foo\n\nBar", "Fizz\n\nBuzz"),
+        ("Another card", "Another back")
+    ]
+    
+    path = tmp_path / "valid_file.fcard"
+    with open(path, "w") as f:
+        f.write(file_contents)
+
+    cards = parser.parse(path)
+
+    assert expected == [(c.front, c.back) for c in cards]
+
+    # Cards have unique IDs
+    assert len(set([c.id for c in cards])) == len(cards)
+
+# Edge cases
+def test_emptyFile(tmp_path):
+    path = tmp_path / "empty.fcard"
+    with open(path, "w") as f:
+        f.write("")
+
+    cards = parser.parse(path)
+    assert cards == []
+
+def test_missingFile(tmp_path):
+    cards = parser.parse(tmp_path / "missing_file.fcard")
+    assert cards == []
+
+def checkException(tmp_path, file_contents):
+    path = tmp_path / "invalid_file.fcard"
+    with open(path, "w") as f:
+        f.write(file_contents)
+
+    with pytest.raises(Exception):
+        cards = parser.parse(path)
+
+def test_doesNotStartWithFront(tmp_path):
+    checkException(tmp_path, "BACK\noops")
+
+def test_frontTwiceInARow(tmp_path):
+    checkException(tmp_path, "FRONT\noops\nFRONT\nbad")
+
+def test_doesNotEndWithBack(tmp_path):
+    checkException(tmp_path, "FRONT\ntest\nBACK\ntest\nFRONT\noops")
-- 
cgit v1.2.3