Heuristics for grabber-continental.com.out (select-one)

Adds specific input types to the field_type.h bitfield to utilize the "select-one" signal in identifying the Country, State, and Credit Card date fields.  Other fields should not classify now if the input type is "select-one".

This is Phase 1 of fixes for the grabber-continental.com.html.  Phase 2 will add lookahead logic for the "BusinessPhone" and "Email Address" issues.

BUG=76299
TEST=FormStructureBrowserTest.DataDrivenHeuristics with test file grabber-continental.com.html

Review URL: http://codereview.chromium.org/7063031

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@86717 0039d316-1c4b-4281-b951-d872f2087c98

chrome/browser/autofill/address_field.cc

@@ -195,7 +195,7 @@ bool AddressField::ParseAddressLines(AutofillScanner* scanner,
         l10n_util::GetStringUTF16(IDS_AUTOFILL_ADDRESS_LINE_1_LABEL_RE);
 
     if (!ParseField(scanner, pattern, &address_field->address1_) &&
-        !ParseFieldSpecifics(scanner, label_pattern, MATCH_LABEL,
+        !ParseFieldSpecifics(scanner, label_pattern, MATCH_LABEL | MATCH_TEXT,
                              &address_field->address1_)) {
       return false;
     }
@@ -214,7 +214,7 @@ bool AddressField::ParseAddressLines(AutofillScanner* scanner,
         l10n_util::GetStringUTF16(IDS_AUTOFILL_ADDRESS_LINE_1_LABEL_RE);
     if (!ParseEmptyLabel(scanner, &address_field->address2_) &&
         !ParseField(scanner, pattern, &address_field->address2_)) {
-      ParseFieldSpecifics(scanner, label_pattern, MATCH_LABEL,
+      ParseFieldSpecifics(scanner, label_pattern, MATCH_LABEL | MATCH_TEXT,
                           &address_field->address2_);
     }
   }
@@ -250,7 +250,8 @@ bool AddressField::ParseCountry(AutofillScanner* scanner,
   else
     pattern = l10n_util::GetStringUTF16(IDS_AUTOFILL_COUNTRY_RE);
 
-  return ParseField(scanner, pattern, &address_field->country_);
+  return ParseFieldSpecifics(scanner, pattern, MATCH_DEFAULT | MATCH_SELECT,
+                             &address_field->country_);
 }
 
 // static
@@ -334,7 +335,8 @@ bool AddressField::ParseState(AutofillScanner* scanner,
   else
     pattern = l10n_util::GetStringUTF16(IDS_AUTOFILL_STATE_RE);
 
-  return ParseField(scanner, pattern, &address_field->state_);
+  return ParseFieldSpecifics(scanner, pattern, MATCH_DEFAULT | MATCH_SELECT,
+                             &address_field->state_);
 }
 
 AddressType AddressField::AddressTypeFromText(const string16 &text) {

chrome/browser/autofill/autofill_resources.grd

@@ -43,7 +43,7 @@
         zip|^-$|post2<!-- pt-BR, pt-PT -->|codpos2
       </message>
       <message name="IDS_AUTOFILL_CITY_RE">
-        city|town<!-- de-DE -->|ort|stadt<!-- en-AU -->|suburb<!-- es -->|ciudad|provincia|localidad|poblacion<!-- fr-FR -->|ville|commune<!-- it-IT -->|localita<!-- ja-JP -->|市区町村<!-- pt-BR, pt-PT -->|cidade<!-- ru -->|Город<!-- zh-CN -->|市<!-- zh-TW -->|分區
+        city|town<!-- de-DE -->|^ort$|stadt<!-- en-AU -->|suburb<!-- es -->|ciudad|provincia|localidad|poblacion<!-- fr-FR -->|ville|commune<!-- it-IT -->|localita<!-- ja-JP -->|市区町村<!-- pt-BR, pt-PT -->|cidade<!-- ru -->|Город<!-- zh-CN -->|市<!-- zh-TW -->|分區
       </message>
       <message name="IDS_AUTOFILL_STATE_RE">
         state|county|region|province<!-- de-DE -->|land<!-- en-UK -->|county|principality<!-- ja-JP -->|都道府県<!-- pt-BR, pt-PT -->|estado|provincia<!-- ru -->|область<!-- zh-CN -->|省<!-- zh-TW -->|地區

chrome/browser/autofill/credit_card_field.cc

@@ -122,13 +122,14 @@ CreditCardField* CreditCardField::Parse(AutofillScanner* scanner,
 
       if ((!credit_card_field->expiration_month_ ||
            credit_card_field->expiration_month_->IsEmpty()) &&
-          ParseField(scanner, pattern, &credit_card_field->expiration_month_)) {
+          ParseFieldSpecifics(scanner, pattern, MATCH_DEFAULT | MATCH_SELECT,
+                              &credit_card_field->expiration_month_)) {
         if (is_ecml)
           pattern = GetEcmlPattern(kEcmlCardExpireYear);
         else
           pattern = l10n_util::GetStringUTF16(IDS_AUTOFILL_EXPIRATION_DATE_RE);
 
-        if (!ParseField(scanner, pattern,
+        if (!ParseFieldSpecifics(scanner, pattern, MATCH_DEFAULT | MATCH_SELECT,
                                  &credit_card_field->expiration_year_)) {
           scanner->Rewind();
           return NULL;
@@ -137,8 +138,10 @@ CreditCardField* CreditCardField::Parse(AutofillScanner* scanner,
       }
     }
 
-    if (ParseField(scanner, GetEcmlPattern(kEcmlCardExpireDay), NULL))
+    if (ParseFieldSpecifics(scanner, GetEcmlPattern(kEcmlCardExpireDay),
+                            MATCH_DEFAULT | MATCH_SELECT, NULL)) {
       continue;
+    }
 
     // Some pages (e.g. ExpediaBilling.html) have a "card description"
     // field; we parse this field but ignore it.

chrome/browser/autofill/email_field.cc

@@ -20,8 +20,10 @@ EmailField* EmailField::Parse(AutofillScanner* scanner, bool is_ecml) {
     pattern = l10n_util::GetStringUTF16(IDS_AUTOFILL_EMAIL_RE);
 
   const AutofillField* field;
-  if (ParseField(scanner, pattern, &field))
+  if (ParseFieldSpecifics(scanner, pattern,
+                          MATCH_DEFAULT | MATCH_EMAIL, &field)) {
     return new EmailField(field);
+  }
 
   return NULL;
 }

chrome/browser/autofill/form_field.cc

@@ -81,6 +81,26 @@ FormField* ParseFormField(AutofillScanner* scanner, bool is_ecml) {
   return NameField::Parse(scanner, is_ecml);
 }
 
+bool IsTextField(const string16& type) {
+  return type == ASCIIToUTF16("text");
+}
+
+bool IsEmailField(const string16& type) {
+  return type == ASCIIToUTF16("email");
+}
+
+bool IsMonthField(const string16& type) {
+  return type == ASCIIToUTF16("month");
+}
+
+bool IsTelephoneField(const string16& type) {
+  return type == ASCIIToUTF16("tel");
+}
+
+bool IsSelectField(const string16& type) {
+  return type == ASCIIToUTF16("select-one");
+}
+
 }  // namespace
 
 // static
@@ -109,7 +129,7 @@ void FormField::ParseFormFields(const std::vector<AutofillField*>& fields,
 bool FormField::ParseField(AutofillScanner* scanner,
                            const string16& pattern,
                            const AutofillField** match) {
-  return ParseFieldSpecifics(scanner, pattern, MATCH_ALL, match);
+  return ParseFieldSpecifics(scanner, pattern, MATCH_DEFAULT, match);
 }
 
 // static
@@ -121,20 +141,31 @@ bool FormField::ParseFieldSpecifics(AutofillScanner* scanner,
     return false;
 
   const AutofillField* field = scanner->Cursor();
-  if (Match(field, pattern, match_type)) {
-    if (match)
-      *match = field;
-    scanner->Advance();
-    return true;
+
+  if ((match_type & MATCH_TEXT) && IsTextField(field->form_control_type))
+    return MatchAndAdvance(scanner, pattern, match_type, match);
+
+  if ((match_type & MATCH_EMAIL) && IsEmailField(field->form_control_type))
+    return MatchAndAdvance(scanner, pattern, match_type, match);
+
+  if ((match_type & MATCH_TELEPHONE) &&
+      IsTelephoneField(field->form_control_type)) {
+    return MatchAndAdvance(scanner, pattern, match_type, match);
   }
 
+  if ((match_type & MATCH_SELECT) && IsSelectField(field->form_control_type))
+    return MatchAndAdvance(scanner, pattern, match_type, match);
+
   return false;
 }
 
 // static
 bool FormField::ParseEmptyLabel(AutofillScanner* scanner,
                                 const AutofillField** match) {
-  return ParseFieldSpecifics(scanner, ASCIIToUTF16("^$"), MATCH_LABEL, match);
+  return ParseFieldSpecifics(scanner,
+                             ASCIIToUTF16("^$"),
+                             MATCH_LABEL | MATCH_ALL_INPUTS,
+                             match);
 }
 
 // static
@@ -148,6 +179,22 @@ bool FormField::AddClassification(const AutofillField* field,
   return map->insert(make_pair(field->unique_name(), type)).second;
 }
 
+// static.
+bool FormField::MatchAndAdvance(AutofillScanner* scanner,
+                                const string16& pattern,
+                                int match_type,
+                                const AutofillField** match) {
+  const AutofillField* field = scanner->Cursor();
+  if (FormField::Match(field, pattern, match_type)) {
+    if (match)
+      *match = field;
+    scanner->Advance();
+    return true;
+  }
+
+  return false;
+}
+
 // static
 bool FormField::Match(const AutofillField* field,
                       const string16& pattern,

chrome/browser/autofill/form_field.h

@@ -32,9 +32,20 @@ class FormField {
  protected:
   // A bit-field used for matching specific parts of a field in question.
   enum MatchType {
+    // Attributes.
     MATCH_LABEL      = 1 << 0,
     MATCH_NAME       = 1 << 1,
-    MATCH_ALL    = MATCH_LABEL | MATCH_NAME
+
+    // Input types.
+    MATCH_TEXT       = 1 << 2,
+    MATCH_EMAIL      = 1 << 3,
+    MATCH_TELEPHONE  = 1 << 4,
+    MATCH_SELECT     = 1 << 5,
+    MATCH_ALL_INPUTS =
+        MATCH_TEXT | MATCH_EMAIL | MATCH_TELEPHONE | MATCH_SELECT,
+
+    // By default match label and name for input/text types.
+    MATCH_DEFAULT    = MATCH_LABEL | MATCH_NAME | MATCH_TEXT,
   };
 
   // Only derived classes may instantiate.
@@ -75,6 +86,15 @@ class FormField {
  private:
   FRIEND_TEST_ALL_PREFIXES(FormFieldTest, Match);
 
+  // Matches |pattern| to the contents of the field at the head of the
+  // |scanner|.
+  // Returns |true| if a match is found according to |match_type|, and |false|
+  // otherwise.
+  static bool MatchAndAdvance(AutofillScanner* scanner,
+                              const string16& pattern,
+                              int match_type,
+                              const AutofillField** match);
+
   // Matches the regular expression |pattern| against the components of |field|
   // as specified in the |match_type| bit field (see |MatchType|).
   static bool Match(const AutofillField* field,

chrome/browser/autofill/phone_field.cc

@@ -269,8 +269,10 @@ bool PhoneField::ParseInternal(PhoneField *phone_field,
     // Attempt to parse according to the next grammar.
     for (; i < arraysize(phone_field_grammars_) &&
          phone_field_grammars_[i].regex != REGEX_SEPARATOR; ++i) {
-      if (!ParseField(scanner,
+      if (!ParseFieldSpecifics(
+              scanner,
               phone_field->GetRegExp(phone_field_grammars_[i].regex),
+              MATCH_DEFAULT | MATCH_TELEPHONE,
               &parsed_fields[phone_field_grammars_[i].phone_part]))
         break;
       if (phone_field_grammars_[i].max_size &&