1 files changed, 47 insertions, 0 deletions
diff --git a/ywalk/parser.py b/ywalk/parser.py
new file mode 100644
index 0000000..40d5f88
--- /dev/null
+++ b/ywalk/parser.py
@@ -0,0 +1,47 @@
+import csv
+
+from ywalk.types import Connection, Mode, Place
+
+FIELD_NAMES = ['Origin', 'Destination', 'Mode', 'Time']
+
+class InputError(Exception):
+    def __init__(self, message, file, line, context):
+        super().__init__(message)
+        self.message = message
+        self.file = file
+        self.line = line
+        self.context = context
+
+    def __str__(self):
+        return f'{self.file}:{self.line}: {self.message}\n {self.context}'
+
+def parse_row(row):
+    for field in FIELD_NAMES:
+        if row[field] is None:
+            raise ValueError(f'Empty field \'{field}\'')
+
+    # This is an enum lookup, not a creation. The Enum class replaces a custom
+    # __new__ after class creation with one that only does lookups.
+    # pylint: disable=no-value-for-parameter
+    mode = Mode(row['Mode'])
+    time = int(row['Time'])
+
+    if time < 0:
+        raise ValueError('Negative travel time')
+
+    return Connection(Place(row['Origin']), Place(row['Destination']), mode, time)
+
+def check_fieldnames(names):
+    for name in FIELD_NAMES:
+        if name not in names:
+            raise ValueError(f'Missing field \'{name}\' in {names}')
+
+def parse_tsv(path):
+    with open(path, encoding='utf-8', newline='') as file:
+        reader = csv.DictReader(file, dialect=csv.excel_tab)
+        check_fieldnames(reader.fieldnames)
+        for line, row in enumerate(reader):
+            try:
+                yield parse_row(row)
+            except ValueError as err:
+                raise InputError(err, path, line + 1, row) from err