form_structure.h revision 7dbb3d5cf0c15f500944d211057644d6a2f37371
1// Copyright 2013 The Chromium Authors. All rights reserved. 2// Use of this source code is governed by a BSD-style license that can be 3// found in the LICENSE file. 4 5#ifndef COMPONENTS_AUTOFILL_CORE_BROWSER_FORM_STRUCTURE_H_ 6#define COMPONENTS_AUTOFILL_CORE_BROWSER_FORM_STRUCTURE_H_ 7 8#include <string> 9#include <vector> 10 11#include "base/gtest_prod_util.h" 12#include "base/memory/scoped_ptr.h" 13#include "base/memory/scoped_vector.h" 14#include "components/autofill/core/browser/autofill_field.h" 15#include "components/autofill/core/browser/autofill_type.h" 16#include "components/autofill/core/browser/field_types.h" 17#include "components/autofill/core/common/web_element_descriptor.h" 18#include "url/gurl.h" 19 20enum RequestMethod { 21 GET, 22 POST 23}; 24 25enum UploadRequired { 26 UPLOAD_NOT_REQUIRED, 27 UPLOAD_REQUIRED, 28 USE_UPLOAD_RATES 29}; 30 31namespace base { 32class TimeTicks; 33} 34 35namespace buzz { 36class XmlElement; 37} 38 39namespace autofill { 40 41class AutofillMetrics; 42 43struct AutocheckoutPageMetaData; 44struct FormData; 45struct FormDataPredictions; 46 47// FormStructure stores a single HTML form together with the values entered 48// in the fields along with additional information needed by Autofill. 49class FormStructure { 50 public: 51 // Whether the form fields should be parsed to match the semantics of plain 52 // ol' Autofill, or of the interactive Autofill dialog. 53 enum ParseTarget { 54 PARSE_FOR_AUTOFILL, 55 PARSE_FOR_AUTOFILL_DIALOG, 56 }; 57 58 FormStructure(const FormData& form, 59 const std::string& autocheckout_url_prefix); 60 virtual ~FormStructure(); 61 62 // Runs several heuristics against the form fields to determine their possible 63 // types. 64 void DetermineHeuristicTypes(const AutofillMetrics& metric_logger); 65 66 // Encodes the XML upload request from this FormStructure. 67 bool EncodeUploadRequest(const FieldTypeSet& available_field_types, 68 bool form_was_autofilled, 69 std::string* encoded_xml) const; 70 71 // Encodes a XML block contains autofill field type from this FormStructure. 72 // This XML will be written VLOG only, never be sent to server. It will 73 // help make FieldAssignments and feed back to autofill server as 74 // experiment data. 75 bool EncodeFieldAssignments(const FieldTypeSet& available_field_types, 76 std::string* encoded_xml) const; 77 78 // Encodes the XML query request for the set of forms. 79 // All fields are returned in one XML. For example, there are three forms, 80 // with 2, 4, and 3 fields. The returned XML would have type info for 9 81 // fields, first two of which would be for the first form, next 4 for the 82 // second, and the rest is for the third. 83 static bool EncodeQueryRequest(const std::vector<FormStructure*>& forms, 84 std::vector<std::string>* encoded_signatures, 85 std::string* encoded_xml); 86 87 // Parses the field types from the server query response. |forms| must be the 88 // same as the one passed to EncodeQueryRequest when constructing the query. 89 static void ParseQueryResponse( 90 const std::string& response_xml, 91 const std::vector<FormStructure*>& forms, 92 autofill::AutocheckoutPageMetaData* page_meta_data, 93 const AutofillMetrics& metric_logger); 94 95 // Fills |forms| with the details from the given |form_structures| and their 96 // fields' predicted types. 97 static void GetFieldTypePredictions( 98 const std::vector<FormStructure*>& form_structures, 99 std::vector<FormDataPredictions>* forms); 100 101 // The unique signature for this form, composed of the target url domain, 102 // the form name, and the form field names in a 64-bit hash. 103 std::string FormSignature() const; 104 105 // Runs a quick heuristic to rule out forms that are obviously not 106 // auto-fillable, like google/yahoo/msn search, etc. The requirement that the 107 // form's method be POST is only applied if |require_method_post| is true. 108 bool IsAutofillable(bool require_method_post) const; 109 110 // Resets |autofill_count_| and counts the number of auto-fillable fields. 111 // This is used when we receive server data for form fields. At that time, 112 // we may have more known fields than just the number of fields we matched 113 // heuristically. 114 void UpdateAutofillCount(); 115 116 // Returns true if this form matches the structural requirements for Autofill. 117 // The requirement that the form's method be POST is only applied if 118 // |require_method_post| is true. 119 bool ShouldBeParsed(bool require_method_post) const; 120 121 // Returns true if we should query the crowdsourcing server to determine this 122 // form's field types. If the form includes author-specified types, this will 123 // return false. 124 bool ShouldBeCrowdsourced() const; 125 126 // Sets the field types and experiment id to be those set for |cached_form|. 127 void UpdateFromCache(const FormStructure& cached_form); 128 129 // Logs quality metrics for |this|, which should be a user-submitted form. 130 // This method should only be called after the possible field types have been 131 // set for each field. |interaction_time| should be a timestamp corresponding 132 // to the user's first interaction with the form. |submission_time| should be 133 // a timestamp corresponding to the form's submission. 134 void LogQualityMetrics(const AutofillMetrics& metric_logger, 135 const base::TimeTicks& load_time, 136 const base::TimeTicks& interaction_time, 137 const base::TimeTicks& submission_time) const; 138 139 // Classifies each field in |fields_| based upon its |autocomplete| attribute, 140 // if the attribute is available. The association is stored into the field's 141 // |heuristic_type|. The exact method of classification depends on 142 // |parse_target|, as the Autofill dialog has slightly different semantics 143 // from regular ol' Autofill. 144 // Fills |found_types| with |true| if the attribute is available and neither 145 // empty nor set to the special values "on" or "off" for at least one field. 146 // Fills |found_sections| with |true| if the attribute specifies a section for 147 // at least one field. 148 void ParseFieldTypesFromAutocompleteAttributes(ParseTarget parse_target, 149 bool* found_types, 150 bool* found_sections); 151 152 const AutofillField* field(size_t index) const; 153 AutofillField* field(size_t index); 154 size_t field_count() const; 155 156 // Returns the number of fields that are able to be autofilled. 157 size_t autofill_count() const { return autofill_count_; } 158 159 // Used for iterating over the fields. 160 std::vector<AutofillField*>::const_iterator begin() const { 161 return fields_.begin(); 162 } 163 std::vector<AutofillField*>::const_iterator end() const { 164 return fields_.end(); 165 } 166 167 const GURL& source_url() const { return source_url_; } 168 169 UploadRequired upload_required() const { return upload_required_; } 170 171 virtual std::string server_experiment_id() const; 172 173 // Returns a FormData containing the data this form structure knows about. 174 // |user_submitted| is currently always false. 175 FormData ToFormData() const; 176 177 bool filled_by_autocheckout() const { return filled_by_autocheckout_; } 178 void set_filled_by_autocheckout(bool filled_by_autocheckout) { 179 filled_by_autocheckout_ = filled_by_autocheckout; 180 } 181 182 bool operator==(const FormData& form) const; 183 bool operator!=(const FormData& form) const; 184 185 private: 186 friend class FormStructureTest; 187 FRIEND_TEST_ALL_PREFIXES(AutofillDownloadTest, QueryAndUploadTest); 188 189 // 64-bit hash of the string - used in FormSignature and unit-tests. 190 static std::string Hash64Bit(const std::string& str); 191 192 enum EncodeRequestType { 193 QUERY, 194 UPLOAD, 195 FIELD_ASSIGNMENTS, 196 }; 197 198 // Adds form info to |encompassing_xml_element|. |request_type| indicates if 199 // it is a query or upload. 200 bool EncodeFormRequest(EncodeRequestType request_type, 201 buzz::XmlElement* encompassing_xml_element) const; 202 203 // Classifies each field in |fields_| into a logical section. 204 // Sections are identified by the heuristic that a logical section should not 205 // include multiple fields of the same autofill type (with some exceptions, as 206 // described in the implementation). Sections are furthermore distinguished 207 // as either credit card or non-credit card sections. 208 // If |has_author_specified_sections| is true, only the second pass -- 209 // distinguishing credit card sections from non-credit card ones -- is made. 210 void IdentifySections(bool has_author_specified_sections); 211 212 bool IsAutocheckoutEnabled() const; 213 214 // Returns true if field should be skipped when talking to Autofill server. 215 bool ShouldSkipField(const FormFieldData& field) const; 216 217 // Returns the minimal number of fillable fields required to start autofill. 218 size_t RequiredFillableFields() const; 219 size_t active_field_count() const; 220 221 // The name of the form. 222 base::string16 form_name_; 223 224 // The source URL. 225 GURL source_url_; 226 227 // The target URL. 228 GURL target_url_; 229 230 // The number of fields able to be auto-filled. 231 size_t autofill_count_; 232 233 // A vector of all the input fields in the form. 234 ScopedVector<AutofillField> fields_; 235 236 // The number of fields counted towards form signature and request to Autofill 237 // server. 238 size_t active_field_count_; 239 240 // The names of the form input elements, that are part of the form signature. 241 // The string starts with "&" and the names are also separated by the "&" 242 // character. E.g.: "&form_input1_name&form_input2_name&...&form_inputN_name" 243 std::string form_signature_field_names_; 244 245 // Whether the server expects us to always upload, never upload, or default 246 // to the stored upload rates. 247 UploadRequired upload_required_; 248 249 // The server experiment corresponding to the server types returned for this 250 // form. 251 std::string server_experiment_id_; 252 253 // GET or POST. 254 RequestMethod method_; 255 256 // Whether the form includes any field types explicitly specified by the site 257 // author, via the |autocompletetype| attribute. 258 bool has_author_specified_types_; 259 260 // The URL prefix matched in autocheckout whitelist. An empty string implies 261 // autocheckout is not enabled for this form. 262 std::string autocheckout_url_prefix_; 263 264 // Whether or not this form was filled by Autocheckout. 265 bool filled_by_autocheckout_; 266 267 DISALLOW_COPY_AND_ASSIGN(FormStructure); 268}; 269 270} // namespace autofill 271 272#endif // COMPONENTS_AUTOFILL_CORE_BROWSER_FORM_STRUCTURE_H_ 273