Doc/lib/liburllib.tex

   1 \section{\module{urllib} ---
   2          Open arbitrary resources by URL}
   3
   4 \declaremodule{standard}{urllib}
   5 \modulesynopsis{Open an arbitrary network resource by URL (requires sockets).}
   6
   7 \index{WWW}
   8 \index{World Wide Web}
   9 \index{URL}
  10
  11
  12 This module provides a high-level interface for fetching data across
  13 the World Wide Web.  In particular, the \function{urlopen()} function
  14 is similar to the built-in function \function{open()}, but accepts
  15 Universal Resource Locators (URLs) instead of filenames.  Some
  16 restrictions apply --- it can only open URLs for reading, and no seek
  17 operations are available.
  18
  19 It defines the following public functions:
  20
  21 \begin{funcdesc}{urlopen}{url\optional{, data\optional{, proxies}}}
  22 Open a network object denoted by a URL for reading.  If the URL does
  23 not have a scheme identifier, or if it has \file{file:} as its scheme
  24 identifier, this opens a local file (without universal newlines);
  25 otherwise it opens a socket to a server somewhere on the network.  If
  26 the connection cannot be made, or if the server returns an error code,
  27 the \exception{IOError} exception is raised.  If all went well, a
  28 file-like object is returned.  This supports the following methods:
  29 \method{read()}, \method{readline()}, \method{readlines()}, \method{fileno()},
  30 \method{close()}, \method{info()} and \method{geturl()}.  It also has
  31 proper support for the iterator protocol.
  32 One caveat: the \method{read()} method, if the size argument is
  33 omitted or negative, may not read until the end of the data stream;
  34 there is no good way to determine that the entire stream from a socket
  35 has been read in the general case.
  36
  37 Except for the \method{info()} and \method{geturl()} methods,
  38 these methods have the same interface as for
  39 file objects --- see section \ref{bltin-file-objects} in this
  40 manual.  (It is not a built-in file object, however, so it can't be
  41 used at those few places where a true built-in file object is
  42 required.)
  43
  44 The \method{info()} method returns an instance of the class
  45 \class{mimetools.Message} containing meta-information associated
  46 with the URL.  When the method is HTTP, these headers are those
  47 returned by the server at the head of the retrieved HTML page
  48 (including Content-Length and Content-Type).  When the method is FTP,
  49 a Content-Length header will be present if (as is now usual) the
  50 server passed back a file length in response to the FTP retrieval
  51 request. A Content-Type header will be present if the MIME type can
  52 be guessed.  When the method is local-file, returned headers will include
  53 a Date representing the file's last-modified time, a Content-Length
  54 giving file size, and a Content-Type containing a guess at the file's
  55 type. See also the description of the
  56 \refmodule{mimetools}\refstmodindex{mimetools} module.
  57
  58 The \method{geturl()} method returns the real URL of the page.  In
  59 some cases, the HTTP server redirects a client to another URL.  The
  60 \function{urlopen()} function handles this transparently, but in some
  61 cases the caller needs to know which URL the client was redirected
  62 to.  The \method{geturl()} method can be used to get at this
  63 redirected URL.
  64
  65 If the \var{url} uses the \file{http:} scheme identifier, the optional
  66 \var{data} argument may be given to specify a \code{POST} request
  67 (normally the request type is \code{GET}).  The \var{data} argument
  68 must be in standard \mimetype{application/x-www-form-urlencoded} format;
  69 see the \function{urlencode()} function below.
  70
  71 The \function{urlopen()} function works transparently with proxies
  72 which do not require authentication.  In a \UNIX{} or Windows
  73 environment, set the \envvar{http_proxy}, \envvar{ftp_proxy} or
  74 \envvar{gopher_proxy} environment variables to a URL that identifies
  75 the proxy server before starting the Python interpreter.  For example
  76 (the \character{\%} is the command prompt):
  77
  78 \begin{verbatim}
  79 % http_proxy="http://www.someproxy.com:3128"
  80 % export http_proxy
  81 % python
  82 ...
  83 \end{verbatim}
  84
  85 In a Windows environment, if no proxy environment variables are set,
  86 proxy settings are obtained from the registry's Internet Settings
  87 section.
  88
  89 In a Macintosh environment, \function{urlopen()} will retrieve proxy
  90 information from Internet\index{Internet Config} Config.
  91
  92 Alternatively, the optional \var{proxies} argument may be used to
  93 explicitly specify proxies.  It must be a dictionary mapping scheme
  94 names to proxy URLs, where an empty dictionary causes no proxies to be
  95 used, and \code{None} (the default value) causes environmental proxy
  96 settings to be used as discussed above.  For example:
  97
  98 \begin{verbatim}
  99 # Use http://www.someproxy.com:3128 for http proxying
 100 proxies = {'http': 'http://www.someproxy.com:3128'}
 101 filehandle = urllib.urlopen(some_url, proxies=proxies)
 102 # Don't use any proxies
 103 filehandle = urllib.urlopen(some_url, proxies={})
 104 # Use proxies from environment - both versions are equivalent
 105 filehandle = urllib.urlopen(some_url, proxies=None)
 106 filehandle = urllib.urlopen(some_url)
 107 \end{verbatim}
 108
 109 The \function{urlopen()} function does not support explicit proxy
 110 specification.  If you need to override environmental proxy settings,
 111 use \class{URLopener}, or a subclass such as \class{FancyURLopener}.
 112
 113 Proxies which require authentication for use are not currently
 114 supported; this is considered an implementation limitation.
 115
 116 \versionchanged[Added the \var{proxies} support]{2.3}
 117 \end{funcdesc}
 118
 119 \begin{funcdesc}{urlretrieve}{url\optional{, filename\optional{,
 120                               reporthook\optional{, data}}}}
 121 Copy a network object denoted by a URL to a local file, if necessary.
 122 If the URL points to a local file, or a valid cached copy of the
 123 object exists, the object is not copied.  Return a tuple
 124 \code{(\var{filename}, \var{headers})} where \var{filename} is the
 125 local file name under which the object can be found, and \var{headers}
 126 is whatever the \method{info()} method of the object returned by
 127 \function{urlopen()} returned (for a remote object, possibly cached).
 128 Exceptions are the same as for \function{urlopen()}.
 129
 130 The second argument, if present, specifies the file location to copy
 131 to (if absent, the location will be a tempfile with a generated name).
 132 The third argument, if present, is a hook function that will be called
 133 once on establishment of the network connection and once after each
 134 block read thereafter.  The hook will be passed three arguments; a
 135 count of blocks transferred so far, a block size in bytes, and the
 136 total size of the file.  The third argument may be \code{-1} on older
 137 FTP servers which do not return a file size in response to a retrieval
 138 request.
 139
 140 If the \var{url} uses the \file{http:} scheme identifier, the optional
 141 \var{data} argument may be given to specify a \code{POST} request
 142 (normally the request type is \code{GET}).  The \var{data} argument
 143 must in standard \mimetype{application/x-www-form-urlencoded} format;
 144 see the \function{urlencode()} function below.
 145
 146 \versionchanged[
 147 \function{urlretrieve()} will raise \exception{ContentTooShortError}
 148 when it detects that the amount of data available
 149 was less than the expected amount (which is the size reported by a
 150 \var{Content-Length} header). This can occur, for example, when the
 151 download is interrupted.
 152
 153 The \var{Content-Length} is treated as a lower bound: if there's more data
 154 to read, urlretrieve reads more data, but if less data is available,
 155 it raises the exception.
 156
 157 You can still retrieve the downloaded data in this case, it is stored
 158 in the \member{content} attribute of the exception instance.
 159
 160 If no \var{Content-Length} header was supplied, urlretrieve can
 161 not check the size of the data it has downloaded, and just returns it.
 162 In this case you just have to assume that the download was successful]{2.5}
 163
 164 \end{funcdesc}
 165
 166 \begin{datadesc}{_urlopener}
 167 The public functions \function{urlopen()} and
 168 \function{urlretrieve()} create an instance of the
 169 \class{FancyURLopener} class and use it to perform their requested
 170 actions.  To override this functionality, programmers can create a
 171 subclass of \class{URLopener} or \class{FancyURLopener}, then assign
 172 an instance of that class to the
 173 \code{urllib._urlopener} variable before calling the desired function.
 174 For example, applications may want to specify a different
 175 \mailheader{User-Agent} header than \class{URLopener} defines.  This
 176 can be accomplished with the following code:
 177
 178 \begin{verbatim}
 179 import urllib
 180
 181 class AppURLopener(urllib.FancyURLopener):
 182     version = "App/1.7"
 183
 184 urllib._urlopener = AppURLopener()
 185 \end{verbatim}
 186 \end{datadesc}
 187
 188 \begin{funcdesc}{urlcleanup}{}
 189 Clear the cache that may have been built up by previous calls to
 190 \function{urlretrieve()}.
 191 \end{funcdesc}
 192
 193 \begin{funcdesc}{quote}{string\optional{, safe}}
 194 Replace special characters in \var{string} using the \samp{\%xx} escape.
 195 Letters, digits, and the characters \character{_.-} are never quoted.
 196 The optional \var{safe} parameter specifies additional characters
 197 that should not be quoted --- its default value is \code{'/'}.
 198
 199 Example: \code{quote('/\~{}connolly/')} yields \code{'/\%7econnolly/'}.
 200 \end{funcdesc}
 201
 202 \begin{funcdesc}{quote_plus}{string\optional{, safe}}
 203 Like \function{quote()}, but also replaces spaces by plus signs, as
 204 required for quoting HTML form values.  Plus signs in the original
 205 string are escaped unless they are included in \var{safe}.  It also
 206 does not have \var{safe} default to \code{'/'}.
 207 \end{funcdesc}
 208
 209 \begin{funcdesc}{unquote}{string}
 210 Replace \samp{\%xx} escapes by their single-character equivalent.
 211
 212 Example: \code{unquote('/\%7Econnolly/')} yields \code{'/\~{}connolly/'}.
 213 \end{funcdesc}
 214
 215 \begin{funcdesc}{unquote_plus}{string}
 216 Like \function{unquote()}, but also replaces plus signs by spaces, as
 217 required for unquoting HTML form values.
 218 \end{funcdesc}
 219
 220 \begin{funcdesc}{urlencode}{query\optional{, doseq}}
 221 Convert a mapping object or a sequence of two-element tuples  to a
 222 ``url-encoded'' string, suitable to pass to
 223 \function{urlopen()} above as the optional \var{data} argument.  This
 224 is useful to pass a dictionary of form fields to a \code{POST}
 225 request.  The resulting string is a series of
 226 \code{\var{key}=\var{value}} pairs separated by \character{\&}
 227 characters, where both \var{key} and \var{value} are quoted using
 228 \function{quote_plus()} above.  If the optional parameter \var{doseq} is
 229 present and evaluates to true, individual \code{\var{key}=\var{value}} pairs
 230 are generated for each element of the sequence.
 231 When a sequence of two-element tuples is used as the \var{query} argument,
 232 the first element of each tuple is a key and the second is a value.  The
 233 order of parameters in the encoded string will match the order of parameter
 234 tuples in the sequence.
 235 The \refmodule{cgi} module provides the functions
 236 \function{parse_qs()} and \function{parse_qsl()} which are used to
 237 parse query strings into Python data structures.
 238 \end{funcdesc}
 239
 240 \begin{funcdesc}{pathname2url}{path}
 241 Convert the pathname \var{path} from the local syntax for a path to
 242 the form used in the path component of a URL.  This does not produce a
 243 complete URL.  The return value will already be quoted using the
 244 \function{quote()} function.
 245 \end{funcdesc}
 246
 247 \begin{funcdesc}{url2pathname}{path}
 248 Convert the path component \var{path} from an encoded URL to the local
 249 syntax for a path.  This does not accept a complete URL.  This
 250 function uses \function{unquote()} to decode \var{path}.
 251 \end{funcdesc}
 252
 253 \begin{classdesc}{URLopener}{\optional{proxies\optional{, **x509}}}
 254 Base class for opening and reading URLs.  Unless you need to support
 255 opening objects using schemes other than \file{http:}, \file{ftp:},
 256 \file{gopher:} or \file{file:}, you probably want to use
 257 \class{FancyURLopener}.
 258
 259 By default, the \class{URLopener} class sends a
 260 \mailheader{User-Agent} header of \samp{urllib/\var{VVV}}, where
 261 \var{VVV} is the \module{urllib} version number.  Applications can
 262 define their own \mailheader{User-Agent} header by subclassing
 263 \class{URLopener} or \class{FancyURLopener} and setting the class
 264 attribute \member{version} to an appropriate string value in the
 265 subclass definition.
 266
 267 The optional \var{proxies} parameter should be a dictionary mapping
 268 scheme names to proxy URLs, where an empty dictionary turns proxies
 269 off completely.  Its default value is \code{None}, in which case
 270 environmental proxy settings will be used if present, as discussed in
 271 the definition of \function{urlopen()}, above.
 272
 273 Additional keyword parameters, collected in \var{x509}, are used for
 274 authentication with the \file{https:} scheme.  The keywords
 275 \var{key_file} and \var{cert_file} are supported; both are needed to
 276 actually retrieve a resource at an \file{https:} URL.
 277 \end{classdesc}
 278
 279 \begin{classdesc}{FancyURLopener}{...}
 280 \class{FancyURLopener} subclasses \class{URLopener} providing default
 281 handling for the following HTTP response codes: 301, 302, 303, 307 and
 282 401.  For the 30x response codes listed above, the
 283 \mailheader{Location} header is used to fetch the actual URL.  For 401
 284 response codes (authentication required), basic HTTP authentication is
 285 performed.  For the 30x response codes, recursion is bounded by the
 286 value of the \var{maxtries} attribute, which defaults to 10.
 287
 288 \note{According to the letter of \rfc{2616}, 301 and 302 responses to
 289   POST requests must not be automatically redirected without
 290   confirmation by the user.  In reality, browsers do allow automatic
 291   redirection of these responses, changing the POST to a GET, and
 292   \module{urllib} reproduces this behaviour.}
 293
 294 The parameters to the constructor are the same as those for
 295 \class{URLopener}.
 296
 297 \note{When performing basic authentication, a
 298 \class{FancyURLopener} instance calls its
 299 \method{prompt_user_passwd()} method.  The default implementation asks
 300 the users for the required information on the controlling terminal.  A
 301 subclass may override this method to support more appropriate behavior
 302 if needed.}
 303 \end{classdesc}
 304
 305 \begin{excclassdesc}{ContentTooShortError}{msg\optional{, content}}
 306 This exception is raised when the \function{urlretrieve()} function
 307 detects that the amount of the downloaded data is less than the
 308 expected amount (given by the \var{Content-Length} header). The
 309 \member{content} attribute stores the downloaded (and supposedly
 310 truncated) data.
 311 \versionadded{2.5}
 312 \end{excclassdesc}
 313
 314 Restrictions:
 315
 316 \begin{itemize}
 317
 318 \item
 319 Currently, only the following protocols are supported: HTTP, (versions
 320 0.9 and 1.0), Gopher (but not Gopher-+), FTP, and local files.
 321 \indexii{HTTP}{protocol}
 322 \indexii{Gopher}{protocol}
 323 \indexii{FTP}{protocol}
 324
 325 \item
 326 The caching feature of \function{urlretrieve()} has been disabled
 327 until I find the time to hack proper processing of Expiration time
 328 headers.
 329
 330 \item
 331 There should be a function to query whether a particular URL is in
 332 the cache.
 333
 334 \item
 335 For backward compatibility, if a URL appears to point to a local file
 336 but the file can't be opened, the URL is re-interpreted using the FTP
 337 protocol.  This can sometimes cause confusing error messages.
 338
 339 \item
 340 The \function{urlopen()} and \function{urlretrieve()} functions can
 341 cause arbitrarily long delays while waiting for a network connection
 342 to be set up.  This means that it is difficult to build an interactive
 343 Web client using these functions without using threads.
 344
 345 \item
 346 The data returned by \function{urlopen()} or \function{urlretrieve()}
 347 is the raw data returned by the server.  This may be binary data
 348 (such as an image), plain text or (for example) HTML\index{HTML}.  The
 349 HTTP\indexii{HTTP}{protocol} protocol provides type information in the
 350 reply header, which can be inspected by looking at the
 351 \mailheader{Content-Type} header.  For the
 352 Gopher\indexii{Gopher}{protocol} protocol, type information is encoded
 353 in the URL; there is currently no easy way to extract it.  If the
 354 returned data is HTML, you can use the module
 355 \refmodule{htmllib}\refstmodindex{htmllib} to parse it.
 356
 357 \item
 358 The code handling the FTP\index{FTP} protocol cannot differentiate
 359 between a file and a directory.  This can lead to unexpected behavior
 360 when attempting to read a URL that points to a file that is not
 361 accessible.  If the URL ends in a \code{/}, it is assumed to refer to
 362 a directory and will be handled accordingly.  But if an attempt to
 363 read a file leads to a 550 error (meaning the URL cannot be found or
 364 is not accessible, often for permission reasons), then the path is
 365 treated as a directory in order to handle the case when a directory is
 366 specified by a URL but the trailing \code{/} has been left off.  This can
 367 cause misleading results when you try to fetch a file whose read
 368 permissions make it inaccessible; the FTP code will try to read it,
 369 fail with a 550 error, and then perform a directory listing for the
 370 unreadable file. If fine-grained control is needed, consider using the
 371 \module{ftplib} module, subclassing \class{FancyURLOpener}, or changing
 372 \var{_urlopener} to meet your needs.
 373
 374 \item
 375 This module does not support the use of proxies which require
 376 authentication.  This may be implemented in the future.
 377
 378 \item
 379 Although the \module{urllib} module contains (undocumented) routines
 380 to parse and unparse URL strings, the recommended interface for URL
 381 manipulation is in module \refmodule{urlparse}\refstmodindex{urlparse}.
 382
 383 \end{itemize}
 384
 385
 386 \subsection{URLopener Objects \label{urlopener-objs}}
 387 \sectionauthor{Skip Montanaro}{skip@mojam.com}
 388
 389 \class{URLopener} and \class{FancyURLopener} objects have the
 390 following attributes.
 391
 392 \begin{methoddesc}[URLopener]{open}{fullurl\optional{, data}}
 393 Open \var{fullurl} using the appropriate protocol.  This method sets
 394 up cache and proxy information, then calls the appropriate open method with
 395 its input arguments.  If the scheme is not recognized,
 396 \method{open_unknown()} is called.  The \var{data} argument
 397 has the same meaning as the \var{data} argument of \function{urlopen()}.
 398 \end{methoddesc}
 399
 400 \begin{methoddesc}[URLopener]{open_unknown}{fullurl\optional{, data}}
 401 Overridable interface to open unknown URL types.
 402 \end{methoddesc}
 403
 404 \begin{methoddesc}[URLopener]{retrieve}{url\optional{,
 405                                         filename\optional{,
 406                                         reporthook\optional{, data}}}}
 407 Retrieves the contents of \var{url} and places it in \var{filename}.  The
 408 return value is a tuple consisting of a local filename and either a
 409 \class{mimetools.Message} object containing the response headers (for remote
 410 URLs) or \code{None} (for local URLs).  The caller must then open and read the
 411 contents of \var{filename}.  If \var{filename} is not given and the URL
 412 refers to a local file, the input filename is returned.  If the URL is
 413 non-local and \var{filename} is not given, the filename is the output of
 414 \function{tempfile.mktemp()} with a suffix that matches the suffix of the last
 415 path component of the input URL.  If \var{reporthook} is given, it must be
 416 a function accepting three numeric parameters.  It will be called after each
 417 chunk of data is read from the network.  \var{reporthook} is ignored for
 418 local URLs.
 419
 420 If the \var{url} uses the \file{http:} scheme identifier, the optional
 421 \var{data} argument may be given to specify a \code{POST} request
 422 (normally the request type is \code{GET}).  The \var{data} argument
 423 must in standard \mimetype{application/x-www-form-urlencoded} format;
 424 see the \function{urlencode()} function below.
 425 \end{methoddesc}
 426
 427 \begin{memberdesc}[URLopener]{version}
 428 Variable that specifies the user agent of the opener object.  To get
 429 \refmodule{urllib} to tell servers that it is a particular user agent,
 430 set this in a subclass as a class variable or in the constructor
 431 before calling the base constructor.
 432 \end{memberdesc}
 433
 434 The \class{FancyURLopener} class offers one additional method that
 435 should be overloaded to provide the appropriate behavior:
 436
 437 \begin{methoddesc}[FancyURLopener]{prompt_user_passwd}{host, realm}
 438 Return information needed to authenticate the user at the given host
 439 in the specified security realm.  The return value should be a tuple,
 440 \code{(\var{user}, \var{password})}, which can be used for basic
 441 authentication.
 442
 443 The implementation prompts for this information on the terminal; an
 444 application should override this method to use an appropriate
 445 interaction model in the local environment.
 446 \end{methoddesc}
 447
 448
 449 \subsection{Examples}
 450 \nodename{Urllib Examples}
 451
 452 Here is an example session that uses the \samp{GET} method to retrieve
 453 a URL containing parameters:
 454
 455 \begin{verbatim}
 456 >>> import urllib
 457 >>> params = urllib.urlencode({'spam': 1, 'eggs': 2, 'bacon': 0})
 458 >>> f = urllib.urlopen("http://www.musi-cal.com/cgi-bin/query?%s" % params)
 459 >>> print f.read()
 460 \end{verbatim}
 461
 462 The following example uses the \samp{POST} method instead:
 463
 464 \begin{verbatim}
 465 >>> import urllib
 466 >>> params = urllib.urlencode({'spam': 1, 'eggs': 2, 'bacon': 0})
 467 >>> f = urllib.urlopen("http://www.musi-cal.com/cgi-bin/query", params)
 468 >>> print f.read()
 469 \end{verbatim}
 470
 471 The following example uses an explicitly specified HTTP proxy,
 472 overriding environment settings:
 473
 474 \begin{verbatim}
 475 >>> import urllib
 476 >>> proxies = {'http': 'http://proxy.example.com:8080/'}
 477 >>> opener = urllib.FancyURLopener(proxies)
 478 >>> f = opener.open("http://www.python.org")
 479 >>> f.read()
 480 \end{verbatim}
 481
 482 The following example uses no proxies at all, overriding environment
 483 settings:
 484
 485 \begin{verbatim}
 486 >>> import urllib
 487 >>> opener = urllib.FancyURLopener({})
 488 >>> f = opener.open("http://www.python.org/")
 489 >>> f.read()
 490 \end{verbatim}