Label scraping for AutoFill....

chrome/browser/autofill/form_structure_unittest.cc

 // Copyright (c) 2010 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "base/scoped_ptr.h"
 #include "base/string_util.h"
 #include "chrome/browser/autofill/form_structure.h"
 #include "googleurl/src/gurl.h"
 #include "testing/gtest/include/gtest/gtest.h"
 #include "third_party/WebKit/WebKit/chromium/public/WebInputElement.h"
@@ skipping 366 matching lines @@
   // Company
   EXPECT_EQ(UNKNOWN_TYPE, form_structure->field(2)->heuristic_type());
   // Address.
   EXPECT_EQ(ADDRESS_HOME_LINE1, form_structure->field(3)->heuristic_type());
   // City.
   EXPECT_EQ(ADDRESS_HOME_CITY, form_structure->field(4)->heuristic_type());
   // Zip.
   EXPECT_EQ(ADDRESS_HOME_ZIP, form_structure->field(5)->heuristic_type());
 }
 
+// Tests a sequence of FormFields where only labels are supplied to heuristics
+// for matching.  This works because FormField labels are matched in the case
+// that input element ids (or |name| fields) are missing.
+TEST(FormStructureTest, HeuristicsLabelsOnly) {
+  scoped_ptr<FormStructure> form_structure;
+  webkit_glue::FormFieldValues values;
+
+  values.method = ASCIIToUTF16("post");
+  values.elements.push_back(webkit_glue::FormField(ASCIIToUTF16("First Name"),
+                                                   string16(),
+                                                   string16(),
+                                                   ASCIIToUTF16("text"),
+                                                   WebInputElement::Text));
+  values.elements.push_back(webkit_glue::FormField(ASCIIToUTF16("Last Name"),
+                                                   string16(),
+                                                   string16(),
+                                                   ASCIIToUTF16("text"),
+                                                   WebInputElement::Text));
+  values.elements.push_back(webkit_glue::FormField(ASCIIToUTF16("EMail"),
+                                                   string16(),
+                                                   string16(),
+                                                   ASCIIToUTF16("text"),
+                                                   WebInputElement::Text));
+  values.elements.push_back(webkit_glue::FormField(ASCIIToUTF16("Phone"),
+                                                   string16(),
+                                                   string16(),
+                                                   ASCIIToUTF16("text"),
+                                                   WebInputElement::Text));
+  values.elements.push_back(webkit_glue::FormField(ASCIIToUTF16("Fax"),
+                                                   string16(),
+                                                   string16(),
+                                                   ASCIIToUTF16("text"),
+                                                   WebInputElement::Text));
+  values.elements.push_back(webkit_glue::FormField(ASCIIToUTF16("Address"),
+                                                   string16(),
+                                                   string16(),
+                                                   ASCIIToUTF16("text"),
+                                                   WebInputElement::Text));
+  values.elements.push_back(webkit_glue::FormField(ASCIIToUTF16("Address"),
+                                                   string16(),
+                                                   string16(),
+                                                   ASCIIToUTF16("text"),
+                                                   WebInputElement::Text));
+  values.elements.push_back(webkit_glue::FormField(ASCIIToUTF16("Zip code"),
+                                                   string16(),
+                                                   string16(),
+                                                   ASCIIToUTF16("text"),
+                                                   WebInputElement::Text));
+  values.elements.push_back(webkit_glue::FormField(string16(),
+                                                   ASCIIToUTF16("Submit"),
+                                                   string16(),
+                                                   ASCIIToUTF16("submit"),
+                                                   WebInputElement::Submit));
+  form_structure.reset(new FormStructure(values));
+  EXPECT_TRUE(form_structure->IsAutoFillable());
+
+  // Expect the correct number of fields.
+  ASSERT_EQ(8UL, form_structure->field_count());
+
+  // Check that heuristics are initialized as UNKNOWN_TYPE.
+  std::vector<AutoFillField*>::const_iterator iter;
+  size_t i;
+  for (iter = form_structure->begin(), i = 0;
+       iter != form_structure->end();
+       ++iter, ++i) {
+    // Expect last element to be NULL.
+    if (i == form_structure->field_count()) {
+      ASSERT_EQ(static_cast<AutoFillField*>(NULL), *iter);
+    } else {
+      ASSERT_NE(static_cast<AutoFillField*>(NULL), *iter);
+      EXPECT_EQ(UNKNOWN_TYPE, (*iter)->heuristic_type());
+    }
+  }
+
+  // Compute heuristic types.
+  form_structure->GetHeuristicAutoFillTypes();
+  ASSERT_EQ(8U, form_structure->field_count());
+
+  // Check that heuristics are no longer UNKNOWN_TYPE.
+  // First name.
+  EXPECT_EQ(NAME_FIRST, form_structure->field(0)->heuristic_type());
+  // Last name.
+  EXPECT_EQ(NAME_LAST, form_structure->field(1)->heuristic_type());
+  // Email.
+  EXPECT_EQ(EMAIL_ADDRESS, form_structure->field(2)->heuristic_type());
+  // Phone.
+  EXPECT_EQ(PHONE_HOME_WHOLE_NUMBER,
+      form_structure->field(3)->heuristic_type());
+  // Fax.  Note, we don't currently match fax.
+  EXPECT_EQ(UNKNOWN_TYPE, form_structure->field(4)->heuristic_type());
+  // Address.
+  EXPECT_EQ(ADDRESS_HOME_LINE1, form_structure->field(5)->heuristic_type());
+  // Address Line 2.
+  EXPECT_EQ(ADDRESS_HOME_LINE2, form_structure->field(6)->heuristic_type());
+  // Zip.
+  EXPECT_EQ(ADDRESS_HOME_ZIP, form_structure->field(7)->heuristic_type());
+}
+
 }  // namespace