|
6043
|
1 // urlwrite and urlread, a curl front-end for octave |
|
|
2 /* |
|
|
3 |
|
|
4 Copyright (C) 2006 Alexander Barth |
|
|
5 |
|
|
6 This file is part of Octave. |
|
|
7 |
|
|
8 Octave is free software; you can redistribute it and/or modify it |
|
|
9 under the terms of the GNU General Public License as published by the |
|
|
10 Free Software Foundation; either version 2, or (at your option) any |
|
|
11 later version. |
|
|
12 |
|
|
13 Octave is distributed in the hope that it will be useful, but WITHOUT |
|
|
14 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or |
|
|
15 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
|
|
16 for more details. |
|
|
17 |
|
|
18 You should have received a copy of the GNU General Public License |
|
|
19 along with Octave; see the file COPYING. If not, write to the Free |
|
|
20 Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA |
|
|
21 02110-1301, USA. |
|
|
22 |
|
|
23 */ |
|
|
24 |
|
|
25 // Author: Alexander Barth <abarth@marine.usf.edu> |
|
|
26 // Adapted-By: jwe |
|
|
27 |
|
|
28 #ifdef HAVE_CONFIG_H |
|
|
29 #include <config.h> |
|
|
30 #endif |
|
|
31 |
|
|
32 #include <string> |
|
|
33 #include <fstream> |
|
|
34 #include <iomanip> |
|
|
35 |
|
|
36 #include "oct-env.h" |
|
|
37 |
|
|
38 #include "defun-dld.h" |
|
|
39 #include "error.h" |
|
|
40 #include "oct-obj.h" |
|
|
41 #include "ov-cell.h" |
|
|
42 #include "pager.h" |
|
|
43 |
|
|
44 #if defined (HAVE_CURL) |
|
|
45 |
|
|
46 #include <curl/curl.h> |
|
|
47 #include <curl/types.h> |
|
|
48 #include <curl/easy.h> |
|
|
49 |
|
|
50 // Write callback function for curl. |
|
|
51 |
|
6986
|
52 static int |
|
6043
|
53 write_data (void *buffer, size_t size, size_t nmemb, void *streamp) |
|
|
54 { |
|
|
55 // *stream is actually an ostream object. |
|
|
56 std::ostream& stream = *(static_cast<std::ostream*> (streamp)); |
|
|
57 stream.write (static_cast<const char*> (buffer), size*nmemb); |
|
|
58 return (stream.fail () ? 0 : size * nmemb); |
|
|
59 } |
|
|
60 |
|
|
61 // Form the query string based on param. |
|
|
62 |
|
6986
|
63 static std::string |
|
6043
|
64 form_query_string (CURL *curl, const Cell& param) |
|
|
65 { |
|
|
66 std::ostringstream query; |
|
|
67 |
|
|
68 for (int i = 0; i < param.numel (); i += 2) |
|
|
69 { |
|
|
70 std::string name = param(i).string_value (); |
|
|
71 std::string text = param(i+1).string_value (); |
|
|
72 |
|
|
73 // Encode strings. |
|
|
74 char *enc_name = curl_easy_escape (curl, name.c_str (), name.length ()); |
|
|
75 char *enc_text = curl_easy_escape (curl, text.c_str (), text.length ()); |
|
|
76 |
|
|
77 query << enc_name << "=" << enc_text; |
|
|
78 |
|
|
79 curl_free (enc_name); |
|
|
80 curl_free (enc_text); |
|
|
81 |
|
|
82 if (i < param.numel()-1) |
|
|
83 query << "&"; |
|
|
84 } |
|
|
85 |
|
|
86 query.flush (); |
|
|
87 |
|
|
88 return query.str (); |
|
|
89 } |
|
|
90 |
|
|
91 // curl front-end |
|
|
92 |
|
6992
|
93 static void |
|
|
94 urlget_cleanup (CURL *curl) |
|
|
95 { |
|
|
96 curl_easy_cleanup (curl); |
|
|
97 curl_global_cleanup (); |
|
|
98 } |
|
|
99 |
|
6986
|
100 static CURLcode |
|
6043
|
101 urlget (const std::string& url, const std::string& method, |
|
|
102 const Cell& param, std::ostream& stream) |
|
|
103 { |
|
|
104 CURL *curl; |
|
|
105 |
|
|
106 curl_global_init(CURL_GLOBAL_DEFAULT); |
|
|
107 |
|
|
108 curl = curl_easy_init(); |
|
|
109 |
|
|
110 if (! curl) |
|
|
111 return CURLE_FAILED_INIT; |
|
|
112 |
|
|
113 // handle paramters of GET or POST request |
|
|
114 |
|
|
115 std::string query_string = form_query_string (curl,param); |
|
|
116 //octave_stdout << "query_string " << query_string << std::endl; |
|
|
117 |
|
|
118 if (method == "get") |
|
|
119 { |
|
|
120 query_string = url + "?" + query_string; |
|
|
121 curl_easy_setopt (curl, CURLOPT_URL, query_string.c_str ()); |
|
|
122 } |
|
|
123 else if (method == "post") |
|
|
124 { |
|
|
125 curl_easy_setopt (curl, CURLOPT_URL, url.c_str ()); |
|
|
126 curl_easy_setopt (curl, CURLOPT_POSTFIELDS, query_string.c_str ()); |
|
|
127 } |
|
|
128 else |
|
7013
|
129 curl_easy_setopt (curl, CURLOPT_URL, url.c_str()); |
|
6043
|
130 |
|
|
131 // Define our callback to get called when there's data to be written. |
|
|
132 curl_easy_setopt (curl, CURLOPT_WRITEFUNCTION, write_data); |
|
|
133 |
|
|
134 // Set a pointer to our struct to pass to the callback. |
|
|
135 curl_easy_setopt (curl, CURLOPT_WRITEDATA, static_cast<void*> (&stream)); |
|
|
136 |
|
6390
|
137 // Follow redirects. |
|
7013
|
138 curl_easy_setopt (curl, CURLOPT_FOLLOWLOCATION, true); |
|
|
139 |
|
|
140 // Don't use EPSV since connecting to sites that don't support it |
|
|
141 // will hang for some time (3 minutes?) before moving on to try PASV |
|
|
142 // instead. |
|
|
143 curl_easy_setopt (curl, CURLOPT_FTP_USE_EPSV, false); |
|
6390
|
144 |
|
6992
|
145 curl_easy_setopt (curl, CURLOPT_NOPROGRESS, true); |
|
6043
|
146 curl_easy_setopt (curl, CURLOPT_PROGRESSDATA, url.c_str ()); |
|
|
147 curl_easy_setopt (curl, CURLOPT_FAILONERROR, true); |
|
|
148 |
|
7013
|
149 // Switch on full protocol/debug output. |
|
|
150 // curl_easy_setopt (curl, CURLOPT_VERBOSE, true); |
|
6043
|
151 |
|
6992
|
152 CURLcode res = CURLE_OK; |
|
|
153 |
|
|
154 // To understand the following, see the definitions of these macros |
|
|
155 // in libcruft/misc/quit.h. The idea is that we call sigsetjmp here |
|
|
156 // then the signal handler calls siglongjmp to get back here |
|
|
157 // immediately. Then we perform some cleanup and throw an interrupt |
|
|
158 // exception which will get us back to the top level, cleaning up |
|
|
159 // any local C++ objects on the stack as we go. |
|
|
160 |
|
|
161 BEGIN_INTERRUPT_IMMEDIATELY_IN_FOREIGN_CODE_1; |
|
|
162 |
|
|
163 // We were interrupted (this code is inside a block that is only |
|
|
164 // called when siglongjmp is called from a signal handler). |
|
6043
|
165 |
|
6992
|
166 // Is there a better error code to use? Maybe it doesn't matter |
|
|
167 // because we are about to throw an execption. |
|
|
168 |
|
|
169 res = CURLE_ABORTED_BY_CALLBACK; |
|
|
170 urlget_cleanup (curl); |
|
|
171 octave_throw_interrupt_exception (); |
|
|
172 |
|
|
173 BEGIN_INTERRUPT_IMMEDIATELY_IN_FOREIGN_CODE_2; |
|
6043
|
174 |
|
6992
|
175 res = curl_easy_perform (curl); |
|
|
176 |
|
|
177 END_INTERRUPT_IMMEDIATELY_IN_FOREIGN_CODE; |
|
|
178 |
|
|
179 // If we are not interuppted, we will end up here, so we still need |
|
|
180 // to clean up. |
|
|
181 |
|
|
182 urlget_cleanup (curl); |
|
6043
|
183 |
|
|
184 return res; |
|
|
185 } |
|
|
186 |
|
|
187 #endif |
|
|
188 |
|
|
189 DEFUN_DLD (urlwrite, args, nargout, |
|
|
190 "-*- texinfo -*-\n\ |
|
|
191 @deftypefn {Loadable Function} {} urlwrite (@var{URL}, @var{localfile})\n\ |
|
|
192 @deftypefnx {Loadable Function} {@var{f} =} urlwrite (@var{url}, @var{localfile})\n\ |
|
|
193 @deftypefnx {Loadable Function} {[@var{f}, @var{success}] =} urlwrite (@var{url}, @var{localfile})\n\ |
|
|
194 @deftypefnx {Loadable Function} {[@var{f}, @var{success}, @var{message}] =} urlwrite (@var{url}, @var{localfile})\n\ |
|
|
195 Download a remote file specified by its @var{URL} and save it as\n\ |
|
|
196 @var{localfile}. For example,\n\ |
|
|
197 \n\ |
|
|
198 @example\n\ |
|
6588
|
199 urlwrite (\"ftp://ftp.octave.org/pub/octave/README\", \"README.txt\");\n\ |
|
6043
|
200 @end example\n\ |
|
|
201 \n\ |
|
|
202 The full path of the downloaded file is returned in @var{f}. The\n\ |
|
|
203 variable @var{success} is 1 if the download was successful,\n\ |
|
|
204 otherwise it is 0 in which case @var{message} contains an error\n\ |
|
|
205 message. If no output argument is specified and if an error occurs,\n\ |
|
6588
|
206 then the error is signaled through Octave's error handling mechanism.\n\ |
|
6043
|
207 \n\ |
|
|
208 This function uses libcurl. Curl supports, among others, the HTTP,\n\ |
|
|
209 FTP and FILE protocols. Username and password may be specified in\n\ |
|
|
210 the URL, for example:\n\ |
|
|
211 \n\ |
|
|
212 @example\n\ |
|
6588
|
213 urlwrite (\"http://username:password@@example.com/file.txt\",\n\ |
|
|
214 \"file.txt\");\n\ |
|
6043
|
215 @end example\n\ |
|
|
216 \n\ |
|
|
217 GET and POST requests can be specified by @var{method} and @var{param}.\n\ |
|
6589
|
218 The parameter @var{method} is either @samp{get} or @samp{post}\n\ |
|
6588
|
219 and @var{param} is a cell array of parameter and value pairs.\n\ |
|
|
220 For example:\n\ |
|
6043
|
221 \n\ |
|
|
222 @example\n\ |
|
6588
|
223 urlwrite (\"http://www.google.com/search\", \"search.html\",\n\ |
|
|
224 \"get\", @{\"query\", \"octave\"@});\n\ |
|
6043
|
225 @end example\n\ |
|
|
226 @seealso{urlread}\n\ |
|
|
227 @end deftypefn") |
|
|
228 { |
|
|
229 octave_value_list retval; |
|
|
230 |
|
|
231 #if defined (HAVE_CURL) |
|
|
232 |
|
|
233 int nargin = args.length (); |
|
|
234 |
|
|
235 // verify arguments |
|
|
236 if (nargin != 2 && nargin != 4) |
|
|
237 { |
|
|
238 print_usage (); |
|
|
239 return retval; |
|
|
240 } |
|
|
241 |
|
|
242 std::string url = args(0).string_value(); |
|
|
243 |
|
|
244 if (error_state) |
|
|
245 { |
|
|
246 error ("urlwrite: url must be a character string"); |
|
|
247 return retval; |
|
|
248 } |
|
|
249 |
|
|
250 // name to store the file if download is succesful |
|
|
251 std::string filename = args(1).string_value(); |
|
|
252 |
|
|
253 if (error_state) |
|
|
254 { |
|
|
255 error ("urlwrite: localfile must be a character string"); |
|
|
256 return retval; |
|
|
257 } |
|
|
258 |
|
|
259 std::string method; |
|
|
260 Cell param; // empty cell array |
|
|
261 |
|
|
262 if (nargin == 4) |
|
|
263 { |
|
|
264 method = args(2).string_value(); |
|
|
265 |
|
|
266 if (error_state) |
|
|
267 { |
|
|
268 error ("urlwrite: method can only be \"get\" or \"post\""); |
|
|
269 return retval; |
|
|
270 } |
|
|
271 |
|
|
272 if (method != "get" && method != "post") |
|
|
273 { |
|
|
274 error ("urlwrite: method can only be \"get\" or \"post\""); |
|
|
275 return retval; |
|
|
276 } |
|
|
277 |
|
|
278 param = args(3).cell_value(); |
|
|
279 |
|
|
280 if (error_state) |
|
|
281 { |
|
|
282 error ("urlwrite: parameters for get and post requests must be given as a cell"); |
|
|
283 return retval; |
|
|
284 } |
|
|
285 |
|
|
286 |
|
|
287 if (param.numel () % 2 == 1 ) |
|
|
288 { |
|
|
289 error ("urlwrite: number of elements in param must be even"); |
|
|
290 return retval; |
|
|
291 } |
|
|
292 } |
|
|
293 |
|
6986
|
294 std::ofstream ofile (filename.c_str(), std::ios::out | std::ios::binary); |
|
6043
|
295 |
|
6986
|
296 if (! ofile.is_open ()) |
|
6043
|
297 { |
|
|
298 error ("urlwrite: unable to open file"); |
|
|
299 return retval; |
|
|
300 } |
|
|
301 |
|
6986
|
302 CURLcode res = urlget (url, method, param, ofile); |
|
6043
|
303 |
|
6986
|
304 ofile.close (); |
|
6043
|
305 |
|
|
306 if (nargout > 0) |
|
|
307 { |
|
|
308 retval(0) = octave_env::make_absolute (filename, octave_env::getcwd ()); |
|
|
309 retval(1) = res == CURLE_OK; |
|
|
310 retval(2) = std::string (res == CURLE_OK ? "" : curl_easy_strerror (res)); |
|
|
311 } |
|
|
312 |
|
6484
|
313 if (nargout < 2 && res != CURLE_OK) |
|
6043
|
314 error ("urlwrite: curl: %s", curl_easy_strerror (res)); |
|
|
315 |
|
|
316 #else |
|
6981
|
317 error ("urlwrite: not available in this version of Octave"); |
|
6043
|
318 #endif |
|
|
319 |
|
|
320 return retval; |
|
|
321 } |
|
|
322 |
|
|
323 DEFUN_DLD (urlread, args, nargout, |
|
|
324 "-*- texinfo -*-\n\ |
|
6549
|
325 @deftypefn {Loadable Function} {@var{s} =} urlread (@var{url})\n\ |
|
6043
|
326 @deftypefnx {Loadable Function} {[@var{s}, @var{success}] =} urlread (@var{url})\n\ |
|
6549
|
327 @deftypefnx {Loadable Function} {[@var{s}, @var{success}, @var{message}] =} urlread (@var{url})\n\ |
|
6547
|
328 @deftypefnx {Loadable Function} {[@dots{}] =} urlread (@var{url}, @var{method}, @var{param})\n\ |
|
6043
|
329 Download a remote file specified by its @var{URL} and return its content\n\ |
|
|
330 in string @var{s}. For example,\n\ |
|
|
331 \n\ |
|
|
332 @example\n\ |
|
6588
|
333 s = urlread (\"ftp://ftp.octave.org/pub/octave/README\");\n\ |
|
6043
|
334 @end example\n\ |
|
|
335 \n\ |
|
|
336 The variable @var{success} is 1 if the download was successful,\n\ |
|
|
337 otherwise it is 0 in which case @var{message} contains an error\n\ |
|
|
338 message. If no output argument is specified and if an error occurs,\n\ |
|
6588
|
339 then the error is signaled through Octave's error handling mechanism.\n\ |
|
6043
|
340 \n\ |
|
|
341 This function uses libcurl. Curl supports, among others, the HTTP,\n\ |
|
|
342 FTP and FILE protocols. Username and password may be specified in the\n\ |
|
|
343 URL. For example,\n\ |
|
|
344 \n\ |
|
|
345 @example\n\ |
|
6588
|
346 s = urlread (\"http://username:password@@example.com/file.txt\");\n\ |
|
6043
|
347 @end example\n\ |
|
|
348 \n\ |
|
|
349 GET and POST requests can be specified by @var{method} and @var{param}.\n\ |
|
6588
|
350 The parameter @var{method} is either @samp{get} or @samp{post}\n\ |
|
|
351 and @var{param} is a cell array of parameter and value pairs.\n\ |
|
6650
|
352 For example,\n\ |
|
6043
|
353 \n\ |
|
|
354 @example\n\ |
|
6588
|
355 s = urlread (\"http://www.google.com/search\", \"get\",\n\ |
|
|
356 @{\"query\", \"octave\"@});\n\ |
|
6043
|
357 @end example\n\ |
|
|
358 @seealso{urlwrite}\n\ |
|
|
359 @end deftypefn") |
|
|
360 { |
|
6588
|
361 // Octave's return value |
|
6043
|
362 octave_value_list retval; |
|
|
363 |
|
|
364 #if defined (HAVE_CURL) |
|
|
365 |
|
|
366 int nargin = args.length (); |
|
|
367 |
|
|
368 // verify arguments |
|
|
369 if (nargin != 1 && nargin != 3) |
|
|
370 { |
|
|
371 print_usage (); |
|
|
372 return retval; |
|
|
373 } |
|
|
374 |
|
|
375 std::string url = args(0).string_value(); |
|
|
376 |
|
|
377 if (error_state) |
|
|
378 { |
|
|
379 error ("urlread: url must be a character string"); |
|
|
380 return retval; |
|
|
381 } |
|
|
382 |
|
|
383 std::string method; |
|
|
384 Cell param; // empty cell array |
|
|
385 |
|
|
386 if (nargin == 3) |
|
|
387 { |
|
|
388 method = args(1).string_value(); |
|
|
389 |
|
|
390 if (error_state) |
|
|
391 { |
|
|
392 error ("urlread: method can only be \"get\" or \"post\""); |
|
|
393 return retval; |
|
|
394 } |
|
|
395 |
|
|
396 if (method != "get" && method != "post") |
|
|
397 { |
|
|
398 error ("urlread: method can only be \"get\" or \"post\""); |
|
|
399 return retval; |
|
|
400 } |
|
|
401 |
|
|
402 param = args(2).cell_value(); |
|
|
403 |
|
|
404 if (error_state) |
|
|
405 { |
|
|
406 error ("urlread: parameters for get and post requests must be given as a cell"); |
|
|
407 return retval; |
|
|
408 } |
|
|
409 |
|
|
410 if (param.numel () % 2 == 1 ) |
|
|
411 { |
|
|
412 error ("urlread: number of elements in param must be even"); |
|
|
413 return retval; |
|
|
414 } |
|
|
415 } |
|
|
416 |
|
6986
|
417 std::ostringstream buf; |
|
6043
|
418 |
|
6986
|
419 CURLcode res = urlget (url, method, param, buf); |
|
6043
|
420 |
|
|
421 if (nargout > 0) |
|
|
422 { |
|
6986
|
423 retval(0) = buf.str (); |
|
6043
|
424 retval(1) = res == CURLE_OK; |
|
6986
|
425 // Return empty string if no error occured. |
|
6043
|
426 retval(2) = std::string (res == CURLE_OK ? "" : curl_easy_strerror (res)); |
|
|
427 } |
|
|
428 |
|
6484
|
429 if (nargout < 2 && res != CURLE_OK) |
|
6043
|
430 error ("urlread: curl: %s", curl_easy_strerror (res)); |
|
|
431 |
|
|
432 #else |
|
6981
|
433 error ("urlread: not available in this version of Octave"); |
|
6043
|
434 #endif |
|
|
435 |
|
|
436 return retval; |
|
|
437 } |
