From 6dc51e73f9d97434280bf7494a0c99d7820fec6a Mon Sep 17 00:00:00 2001
From: Will Woods <wwoods@redhat.com>
Date: Jun 05 2020 17:55:35 +0000
Subject: countme.ItemReader: add .fields and ._get_fields()


Refactor the field-reading / field-check from ItemReader._get_reader()
to make things a little clearer, and add a `.fields` property so callers
can examine what fields are in the thing they're reading.

---

diff --git a/countme/__init__.py b/countme/__init__.py
index 1e82f1e..e3504f8 100644
--- a/countme/__init__.py
+++ b/countme/__init__.py
@@ -334,15 +334,21 @@ class ItemReader:
         self._itemtuple = itemtuple
         self._itemfields = itemtuple._fields
         self._itemfactory = itemtuple._make
-        self._filefields = None
         self._get_reader(**kwargs)
-        if not self._filefields:
+        filefields = self._get_fields()
+        if not filefields:
             raise ReaderError("no field names found")
-        if self._filefields != self._itemfields:
-            raise ReaderError(f"field mismatch: expected {self._itemfields}, got {self._filefields}")
+        if filefields != self._itemfields:
+            raise ReaderError(f"field mismatch: expected {self._itemfields}, got {filefields}")
+    @property
+    def fields(self):
+        return self._itemfields
     def _get_reader(self):
-        '''Set up the ItemReader.
-        Should set self._filefields to a tuple of the fields found in fp.'''
+        '''Set up the ItemReader.'''
+        raise NotImplementedError
+    def _get_fields(self):
+        '''Called immediately after _get_reader().
+        Should return a tuple of the fieldnames found in self._fp.'''
         raise NotImplementedError
     def _iter_rows(self):
         '''Return an iterator/generator that produces a row for each item.'''
@@ -355,11 +361,13 @@ class CSVReader(ItemReader):
     def _get_reader(self, **kwargs):
         import csv
         self._reader = csv.reader(self._fp)
-        self._filefields = tuple(next(self._reader))
-        # If we have numbers in our fieldnames, probably there was no header
-        if any(name.isnumeric() for name in self._filefields):
-            header = ','.join(fields)
-            raise ReaderError(f"header bad/missing, got: {header}")
+    def _get_fields(self):
+        filefields = tuple(next(self._reader))
+        # Sanity check: if any field is a number, this isn't a header
+        if any(name.isnumeric() for name in filefields):
+            header = ','.join(filefields)
+            raise ReaderError(f"header bad/missing: expected {self._itemfields}, got {header!r}")
+        return filefields
     def _iter_rows(self):
         return self._reader
 
@@ -372,12 +380,10 @@ class SQLiteReader(ItemReader):
         # TODO: self._con.set_progress_handler(handler, call_interval)
         self._cur = self._con.cursor()
         self._tablename = tablename
-        if False and sqlite3.sqlite_version_info >= (3,16,0):
-            fields_sql = f"SELECT name FROM pragma_table_info(?)"
-            self._filefields = tuple(r[0] for r in self._cur.execute(fields_sql, (tablename,)))
-        else:
-            fields_sql = f"PRAGMA table_info('{tablename}')"
-            self._filefields = tuple(r[1] for r in self._cur.execute(fields_sql))
+    def _get_fields(self):
+        fields_sql = f"PRAGMA table_info('{self._tablename}')"
+        filefields = tuple(r[1] for r in self._cur.execute(fields_sql))
+        return filefields
     def _iter_rows(self):
         fields = ",".join(self._itemfields)
         return self._cur.execute(f"SELECT {fields} FROM {self._tablename}")