Class: Csvlint::Validator
- Inherits:
-
Object
- Object
- Csvlint::Validator
- Includes:
- ErrorCollector
- Defined in:
- lib/csvlint/validate.rb
Defined Under Namespace
Classes: LineCSV
Constant Summary collapse
- ERROR_MATCHERS =
{ "Missing or stray quote" => :stray_quote, "Illegal quoting" => :whitespace, "Unclosed quoted field" => :unclosed_quote, "Unquoted fields do not allow \\r or \\n" => :line_breaks, }
Instance Attribute Summary collapse
-
#content_type ⇒ Object
readonly
Returns the value of attribute content_type.
-
#csv_header ⇒ Object
readonly
Returns the value of attribute csv_header.
-
#current_line ⇒ Object
readonly
Returns the value of attribute current_line.
-
#data ⇒ Object
readonly
Returns the value of attribute data.
-
#dialect ⇒ Object
readonly
Returns the value of attribute dialect.
-
#encoding ⇒ Object
readonly
Returns the value of attribute encoding.
-
#extension ⇒ Object
readonly
Returns the value of attribute extension.
-
#headers ⇒ Object
readonly
Returns the value of attribute headers.
-
#link_headers ⇒ Object
readonly
Returns the value of attribute link_headers.
-
#schema ⇒ Object
readonly
Returns the value of attribute schema.
Attributes included from ErrorCollector
#errors, #info_messages, #warnings
Instance Method Summary collapse
- #build_exception_messages(csvException, errChars, lineNo) ⇒ Object
- #build_formats(row) ⇒ Object
- #build_linebreak_error ⇒ Object
- #check_consistency ⇒ Object
- #check_foreign_keys ⇒ Object
- #check_mixed_linebreaks ⇒ Object
- #dialect_to_csv_options(dialect) ⇒ Object
- #fetch_error(error) ⇒ Object
- #finish ⇒ Object
- #header? ⇒ Boolean
-
#initialize(source, dialect = {}, schema = nil, options = {}) ⇒ Validator
constructor
A new instance of Validator.
- #line_breaks ⇒ Object
- #line_breaks_reported? ⇒ Boolean
- #locate_schema ⇒ Object
-
#parse_contents(stream, line = nil) ⇒ Object
analyses the provided csv and builds errors, warnings and info messages.
- #parse_line(line) ⇒ Object
- #report_line_breaks(line_no = nil) ⇒ Object
- #row_count ⇒ Object
- #set_dialect ⇒ Object
- #validate ⇒ Object
- #validate_encoding ⇒ Object
- #validate_header(header) ⇒ Object
- #validate_line(input = nil, index = nil) ⇒ Object
- #validate_metadata ⇒ Object
- #validate_stream ⇒ Object
- #validate_url ⇒ Object
Methods included from ErrorCollector
#build_errors, #build_info_messages, #build_warnings, #reset, #valid?
Constructor Details
#initialize(source, dialect = {}, schema = nil, options = {}) ⇒ Validator
Returns a new instance of Validator.
62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 |
# File 'lib/csvlint/validate.rb', line 62 def initialize(source, dialect = {}, schema = nil, = {}) reset @source = source @formats = [] @schema = schema @dialect = dialect @csv_header = true @headers = {} @lambda = [:lambda] @leading = "" @limit_lines = [:limit_lines] @extension = parse_extension(source) unless @source.nil? @expected_columns = 0 @col_counts = [] @line_breaks = [] @errors += @schema.errors unless @schema.nil? @warnings += @schema.warnings unless @schema.nil? @data = [] # it may be advisable to flush this on init? validate end |
Instance Attribute Details
#content_type ⇒ Object (readonly)
Returns the value of attribute content_type.
53 54 55 |
# File 'lib/csvlint/validate.rb', line 53 def content_type @content_type end |
#csv_header ⇒ Object (readonly)
Returns the value of attribute csv_header.
53 54 55 |
# File 'lib/csvlint/validate.rb', line 53 def csv_header @csv_header end |
#current_line ⇒ Object (readonly)
Returns the value of attribute current_line.
53 54 55 |
# File 'lib/csvlint/validate.rb', line 53 def current_line @current_line end |
#data ⇒ Object (readonly)
Returns the value of attribute data.
53 54 55 |
# File 'lib/csvlint/validate.rb', line 53 def data @data end |
#dialect ⇒ Object (readonly)
Returns the value of attribute dialect.
53 54 55 |
# File 'lib/csvlint/validate.rb', line 53 def dialect @dialect end |
#encoding ⇒ Object (readonly)
Returns the value of attribute encoding.
53 54 55 |
# File 'lib/csvlint/validate.rb', line 53 def encoding @encoding end |
#extension ⇒ Object (readonly)
Returns the value of attribute extension.
53 54 55 |
# File 'lib/csvlint/validate.rb', line 53 def extension @extension end |
#headers ⇒ Object (readonly)
Returns the value of attribute headers.
53 54 55 |
# File 'lib/csvlint/validate.rb', line 53 def headers @headers end |
#link_headers ⇒ Object (readonly)
Returns the value of attribute link_headers.
53 54 55 |
# File 'lib/csvlint/validate.rb', line 53 def link_headers @link_headers end |
#schema ⇒ Object (readonly)
Returns the value of attribute schema.
53 54 55 |
# File 'lib/csvlint/validate.rb', line 53 def schema @schema end |
Instance Method Details
#build_exception_messages(csvException, errChars, lineNo) ⇒ Object
335 336 337 338 339 340 341 342 343 344 |
# File 'lib/csvlint/validate.rb', line 335 def (csvException, errChars, lineNo) #TODO 1 - this is a change in logic, rather than straight refactor of previous error building, however original logic is bonkers #TODO 2 - using .kind_of? is a very ugly fix here and it meant to work around instances where :auto symbol is preserved in @csv_options type = fetch_error(csvException) if !@csv_options[:row_sep].kind_of?(Symbol) && [:unclosed_quote,:stray_quote].include?(type) && !@input.match(@csv_options[:row_sep]) build_linebreak_error else build_errors(type, :structure, lineNo, nil, errChars) end end |
#build_formats(row) ⇒ Object
387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 |
# File 'lib/csvlint/validate.rb', line 387 def build_formats(row) row.each_with_index do |col, i| next if col.nil? || col.empty? @formats[i] ||= Hash.new(0) format = if col.strip[FORMATS[:numeric]] :numeric elsif uri?(col) :uri elsif possible_date?(col) date_formats(col) else :string end @formats[i][format] += 1 end end |
#build_linebreak_error ⇒ Object
346 347 348 |
# File 'lib/csvlint/validate.rb', line 346 def build_linebreak_error build_errors(:line_breaks, :structure) unless @errors.any? { |e| e.type == :line_breaks } end |
#check_consistency ⇒ Object
407 408 409 410 411 412 413 414 415 416 |
# File 'lib/csvlint/validate.rb', line 407 def check_consistency @formats.each_with_index do |format,i| if format total = format.values.reduce(:+).to_f if format.none?{|_,count| count / total >= 0.9} build_warnings(:inconsistent_values, :schema, nil, i + 1) end end end end |
#check_foreign_keys ⇒ Object
418 419 420 421 422 423 424 |
# File 'lib/csvlint/validate.rb', line 418 def check_foreign_keys if @schema.instance_of? Csvlint::Csvw::TableGroup @schema.validate_foreign_keys @errors += @schema.errors @warnings += @schema.warnings end end |
#check_mixed_linebreaks ⇒ Object
319 320 321 |
# File 'lib/csvlint/validate.rb', line 319 def check_mixed_linebreaks build_linebreak_error if @line_breaks.uniq.count > 1 end |
#dialect_to_csv_options(dialect) ⇒ Object
375 376 377 378 379 380 381 382 383 384 385 |
# File 'lib/csvlint/validate.rb', line 375 def (dialect) skipinitialspace = dialect["skipInitialSpace"] || true delimiter = dialect["delimiter"] delimiter = delimiter + " " if !skipinitialspace return { :col_sep => delimiter, :row_sep => dialect["lineTerminator"], :quote_char => dialect["quoteChar"], :skip_blanks => false } end |
#fetch_error(error) ⇒ Object
369 370 371 372 373 |
# File 'lib/csvlint/validate.rb', line 369 def fetch_error(error) e = error..match(/^(.+?)(?: [io]n)? \(?line \d+\)?\.?$/i) = e[1] rescue nil ERROR_MATCHERS.fetch(, :unknown_error) end |
#finish ⇒ Object
210 211 212 213 214 215 216 217 218 219 220 221 |
# File 'lib/csvlint/validate.rb', line 210 def finish sum = @col_counts.inject(:+) unless sum.nil? build_warnings(:title_row, :structure) if @col_counts.first < (sum / @col_counts.size.to_f) end # return expected_columns to calling class build_warnings(:check_options, :structure) if @expected_columns == 1 check_consistency check_foreign_keys check_mixed_linebreaks validate_encoding end |
#header? ⇒ Boolean
266 267 268 |
# File 'lib/csvlint/validate.rb', line 266 def header? @csv_header && @dialect["header"] end |
#line_breaks ⇒ Object
323 324 325 326 327 328 329 |
# File 'lib/csvlint/validate.rb', line 323 def line_breaks if @line_breaks.uniq.count > 1 :mixed else @line_breaks.uniq.first end end |
#line_breaks_reported? ⇒ Boolean
282 283 284 |
# File 'lib/csvlint/validate.rb', line 282 def line_breaks_reported? @line_breaks_reported === true end |
#locate_schema ⇒ Object
426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 |
# File 'lib/csvlint/validate.rb', line 426 def locate_schema @source_url = nil warn_if_unsuccessful = false case @source when StringIO return when File @source_url = "file:#{File.(@source)}" else @source_url = @source end unless @schema.nil? if @schema.tables[@source_url] return else @schema = nil end end link_schema = nil @schema = link_schema if link_schema paths = [] if @source_url =~ /^http(s)?/ begin well_known_uri = URI.join(@source_url, "/.well-known/csvm") well_known = open(well_known_uri).read # TODO rescue OpenURI::HTTPError, URI::BadURIError end end paths = ["{+url}-metadata.json", "csv-metadata.json"] if paths.empty? paths.each do |template| begin template = URITemplate.new(template) path = template.('url' => @source_url) url = URI.join(@source_url, path) url = File.new(url.to_s.sub(/^file:/, "")) if url.to_s =~ /^file:/ schema = Schema.load_from_json(url) if schema.instance_of? Csvlint::Csvw::TableGroup if schema.tables[@source_url] @schema = schema else warn_if_unsuccessful = true build_warnings(:schema_mismatch, :context, nil, nil, @source_url, schema) end end rescue Errno::ENOENT rescue OpenURI::HTTPError, URI::BadURIError, ArgumentError rescue => e STDERR.puts e.class STDERR.puts e. STDERR.puts e.backtrace raise e end end build_warnings(:schema_mismatch, :context, nil, nil, @source_url, schema) if warn_if_unsuccessful @schema = nil end |
#parse_contents(stream, line = nil) ⇒ Object
analyses the provided csv and builds errors, warnings and info messages
172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 |
# File 'lib/csvlint/validate.rb', line 172 def parse_contents(stream, line = nil) # parse_contents will parse one line and apply headers, formats methods and error handle as appropriate current_line = line.present? ? line : 1 all_errors = [] @csv_options[:encoding] = @encoding begin row = LineCSV.parse_line(stream, @csv_options) rescue LineCSV::MalformedCSVError => e (e, stream, current_line) end @data << row if row if current_line <= 1 && @csv_header # this conditional should be refactored somewhere row = row.reject { |col| col.nil? || col.empty? } validate_header(row) @col_counts << row.size else build_formats(row) @col_counts << row.reject { |col| col.nil? || col.empty? }.size @expected_columns = row.size unless @expected_columns != 0 build_errors(:blank_rows, :structure, current_line, nil, stream.to_s) if row.reject { |c| c.nil? || c.empty? }.size == 0 # Builds errors and warnings related to the provided schema file if @schema @schema.validate_row(row, current_line, all_errors, @source) @errors += @schema.errors all_errors += @schema.errors @warnings += @schema.warnings else build_errors(:ragged_rows, :structure, current_line, nil, stream.to_s) if !row.empty? && row.size != @expected_columns end end end end |
#parse_line(line) ⇒ Object
136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 |
# File 'lib/csvlint/validate.rb', line 136 def parse_line(line) line = @leading + line # Check if the last line is a line break - in which case it's a full line if line[-1, 1].include?("\n") # If the number of quotes is odd, the linebreak is inside some quotes if line.count(@dialect["quoteChar"]).odd? @leading = line else validate_line(line, @current_line) @leading = "" @current_line = @current_line+1 end else # If it's not a full line, then prepare to add it to the beginning of the next chunk @leading = line end rescue ArgumentError => ae build_errors(:invalid_encoding, :structure, @current_line, nil, @current_line) unless @reported_invalid_encoding @current_line = @current_line+1 @reported_invalid_encoding = true end |
#report_line_breaks(line_no = nil) ⇒ Object
270 271 272 273 274 275 276 277 278 279 280 |
# File 'lib/csvlint/validate.rb', line 270 def report_line_breaks(line_no=nil) return unless @input[-1, 1].include?("\n") # Return straight away if there's no newline character - i.e. we're on the last line line_break = get_line_break(@input) @line_breaks << line_break unless line_breaks_reported? if line_break != "\r\n" (:nonrfc_line_breaks, :structure, line_no) @line_breaks_reported = true end end end |
#row_count ⇒ Object
331 332 333 |
# File 'lib/csvlint/validate.rb', line 331 def row_count data.count end |
#set_dialect ⇒ Object
286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 |
# File 'lib/csvlint/validate.rb', line 286 def set_dialect @assumed_header = @dialect["header"].nil? @supplied_dialect = @dialect != {} begin schema_dialect = @schema.tables[@source_url].dialect || {} rescue schema_dialect = {} end @dialect = { "header" => true, "delimiter" => ",", "skipInitialSpace" => true, "lineTerminator" => :auto, "quoteChar" => '"', "trim" => :true }.merge(schema_dialect).merge(@dialect || {}) @csv_header = @csv_header && @dialect["header"] @csv_options = (@dialect) end |
#validate ⇒ Object
88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 |
# File 'lib/csvlint/validate.rb', line 88 def validate if @extension =~ /.xls(x)?/ build_warnings(:excel, :context) return end locate_schema unless @schema.instance_of?(Csvlint::Schema) set_dialect if @source.class == String validate_url else validate_stream end finish end |
#validate_encoding ⇒ Object
308 309 310 311 312 313 314 315 316 317 |
# File 'lib/csvlint/validate.rb', line 308 def validate_encoding if @headers["content-type"] if @headers["content-type"] !~ /charset=/ build_warnings(:no_encoding, :context) elsif @headers["content-type"] !~ /charset=utf-8/i build_warnings(:encoding, :context) end end build_warnings(:encoding, :context) if @encoding != "UTF-8" end |
#validate_header(header) ⇒ Object
350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 |
# File 'lib/csvlint/validate.rb', line 350 def validate_header(header) names = Set.new header.map{|h| h.strip! } if @dialect["trim"] == :true header.each_with_index do |name,i| build_warnings(:empty_column_name, :schema, nil, i+1) if name == "" if names.include?(name) build_warnings(:duplicate_column_name, :schema, nil, i+1) else names << name end end if @schema @schema.validate_header(header, @source) @errors += @schema.errors @warnings += @schema.warnings end return valid? end |
#validate_line(input = nil, index = nil) ⇒ Object
158 159 160 161 162 163 164 165 166 167 168 169 |
# File 'lib/csvlint/validate.rb', line 158 def validate_line(input = nil, index = nil) @input = input single_col = false line = index.present? ? index : 0 @encoding = input.encoding.to_s report_line_breaks(line) parse_contents(input, line) @lambda.call(self) unless @lambda.nil? rescue ArgumentError => ae build_errors(:invalid_encoding, :structure, @current_line, nil, index) unless @reported_invalid_encoding @reported_invalid_encoding = true end |
#validate_metadata ⇒ Object
223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 |
# File 'lib/csvlint/validate.rb', line 223 def assumed_header = !@supplied_dialect unless @headers.empty? if @headers["content-type"] =~ /text\/csv/ @csv_header = @csv_header && true assumed_header = @assumed_header.present? end if @headers["content-type"] =~ /header=(present|absent)/ @csv_header = true if $1 == "present" @csv_header = false if $1 == "absent" assumed_header = false end build_warnings(:no_content_type, :context) if @content_type == nil build_errors(:wrong_content_type, :context) unless (@content_type && @content_type =~ /text\/csv/) end @header_processed = true (:assumed_header, :structure) if assumed_header @link_headers = @headers["link"].split(",") rescue nil @link_headers.each do |link_header| match = LINK_HEADER_REGEXP.match(link_header) uri = match["uri"].gsub(/(^\<|\>$)/, "") rescue nil rel = match["rel-relationship"].gsub(/(^\"|\"$)/, "") rescue nil param = match["param"] param_value = match["param-value"].gsub(/(^\"|\"$)/, "") rescue nil if rel == "describedby" && param == "type" && ["application/csvm+json", "application/ld+json", "application/json"].include?(param_value) begin url = URI.join(@source_url, uri) schema = Schema.load_from_json(url) if schema.instance_of? Csvlint::Csvw::TableGroup if schema.tables[@source_url] link_schema = schema else warn_if_unsuccessful = true build_warnings(:schema_mismatch, :context, nil, nil, @source_url, schema) end end rescue OpenURI::HTTPError end end end if @link_headers end |
#validate_stream ⇒ Object
105 106 107 108 109 110 111 112 |
# File 'lib/csvlint/validate.rb', line 105 def validate_stream @current_line = 1 @source.each_line do |line| break if line_limit_reached? parse_line(line) end validate_line(@leading, @current_line) unless @leading == "" end |
#validate_url ⇒ Object
114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 |
# File 'lib/csvlint/validate.rb', line 114 def validate_url @current_line = 1 request = Typhoeus::Request.new(@source, followlocation: true) request.on_headers do |response| @headers = response.headers || {} @content_type = response.headers["content-type"] rescue nil @response_code = response.code return build_errors(:not_found) if response.code == 404 end request.on_body do |chunk| io = StringIO.new(chunk) io.each_line do |line| break if line_limit_reached? parse_line(line) end end request.run # Validate the last line too validate_line(@leading, @current_line) unless @leading == "" end |