extras/find-fixme/fixme-finder.c

/***************************************
 OSM planet file fixme finder.

 Part of the Routino routing software.
 ******************/ /******************
 This file Copyright 2008-2013 Andrew M. Bishop

 This program is free software: you can redistribute it and/or modify
 it under the terms of the GNU Affero General Public License as published by
 the Free Software Foundation, either version 3 of the License, or
 (at your option) any later version.

 This program is distributed in the hope that it will be useful,
 but WITHOUT ANY WARRANTY; without even the implied warranty of
 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 GNU Affero General Public License for more details.

 You should have received a copy of the GNU Affero General Public License
 along with this program.  If not, see <http://www.gnu.org/licenses/>.
 ***************************************/


#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <string.h>
#include <errno.h>
#include <sys/time.h>

#include "types.h"
#include "ways.h"

#include "typesx.h"
#include "nodesx.h"
#include "waysx.h"
#include "relationsx.h"

#include "files.h"
#include "logging.h"
#include "errorlogx.h"
#include "functions.h"
#include "osmparser.h"
#include "tagging.h"
#include "uncompress.h"


/* Global variables */

/*+ The name of the temporary directory. +*/
char *option_tmpdirname=NULL;

/*+ The amount of RAM to use for filesorting. +*/
size_t option_filesort_ramsize=0;

/*+ The number of threads to use for filesorting. +*/
int option_filesort_threads=1;


/* Local functions */

static void print_usage(int detail,const char *argerr,const char *err);


/*++++++++++++++++++++++++++++++++++++++
  The main program for the find-fixme.
  ++++++++++++++++++++++++++++++++++++++*/

int main(int argc,char** argv)
{
 struct timeval start_time;
 NodesX     *OSMNodes;
 WaysX      *OSMWays;
 RelationsX *OSMRelations;
 ErrorLogsX *OSMErrorLogs;
 char       *dirname=NULL,*prefix=NULL,*tagging="fixme.xml",*errorlog="fixme.log";
 int         option_keep=1;
 int         option_filenames=0;
 int         arg;

 gettimeofday(&start_time,NULL);

 /* Parse the command line arguments */

 for(arg=1;arg<argc;arg++)
   {
    if(!strcmp(argv[arg],"--help"))
       print_usage(1,NULL,NULL);
    else if(!strncmp(argv[arg],"--dir=",6))
       dirname=&argv[arg][6];
    else if(!strncmp(argv[arg],"--sort-ram-size=",16))
       option_filesort_ramsize=atoi(&argv[arg][16]);
#if defined(USE_PTHREADS) && USE_PTHREADS
    else if(!strncmp(argv[arg],"--sort-threads=",15))
       option_filesort_threads=atoi(&argv[arg][15]);
#endif
    else if(!strncmp(argv[arg],"--tmpdir=",9))
       option_tmpdirname=&argv[arg][9];
    else if(!strncmp(argv[arg],"--tagging=",10))
       tagging=&argv[arg][10];
    else if(!strcmp(argv[arg],"--loggable"))
       option_loggable=1;
    else if(!strcmp(argv[arg],"--logtime"))
       option_logtime=1;
    else if(argv[arg][0]=='-' && argv[arg][1]=='-')
       print_usage(0,argv[arg],NULL);
    else
       option_filenames++;
   }

 /* Check the specified command line options */

 if(!option_filesort_ramsize)
   {
#if SLIM
    option_filesort_ramsize=64*1024*1024;
#else
    option_filesort_ramsize=256*1024*1024;
#endif
   }
 else
    option_filesort_ramsize*=1024*1024;

 if(!option_tmpdirname)
   {
    if(!dirname)
       option_tmpdirname=".";
    else
       option_tmpdirname=dirname;
   }

 if(tagging)
   {
    if(!ExistsFile(tagging))
      {
       fprintf(stderr,"Error: The '--tagging' option specifies a file that does not exist.\n");
       exit(EXIT_FAILURE);
      }
   }
 else
   {
    if(ExistsFile(FileName(dirname,prefix,"tagging.xml")))
       tagging=FileName(dirname,prefix,"tagging.xml");
    else
      {
       fprintf(stderr,"Error: The '--tagging' option was not used and the default 'tagging.xml' does not exist.\n");
       exit(EXIT_FAILURE);
      }
   }

 if(ParseXMLTaggingRules(tagging))
   {
    fprintf(stderr,"Error: Cannot read the tagging rules in the file '%s'.\n",tagging);
    exit(EXIT_FAILURE);
   }

 /* Create new node, segment, way and relation variables */

 OSMNodes=NewNodeList(0,0);

 OSMWays=NewWayList(0,0);

 OSMRelations=NewRelationList(0,0);

 /* Create the error log file */

 if(errorlog)
    open_errorlog(FileName(dirname,prefix,errorlog),0,option_keep);

 /* Parse the file */

 for(arg=1;arg<argc;arg++)
   {
    int fd;
    char *filename,*p;

    if(argv[arg][0]=='-' && argv[arg][1]=='-')
       continue;

    filename=strcpy(malloc(strlen(argv[arg])+1),argv[arg]);

    fd=ReOpenFile(filename);

    if((p=strstr(filename,".bz2")) && !strcmp(p,".bz2"))
      {
       fd=Uncompress_Bzip2(fd);
       *p=0;
      }

    if((p=strstr(filename,".gz")) && !strcmp(p,".gz"))
      {
       fd=Uncompress_Gzip(fd);
       *p=0;
      }

    printf("\nParse OSM Data [%s]\n==============\n\n",filename);
    fflush(stdout);

    if((p=strstr(filename,".pbf")) && !strcmp(p,".pbf"))
      {
       if(ParsePBFFile(fd,OSMNodes,OSMWays,OSMRelations))
          exit(EXIT_FAILURE);
      }
    else if((p=strstr(filename,".o5m")) && !strcmp(p,".o5m"))
      {
       if(ParseO5MFile(fd,OSMNodes,OSMWays,OSMRelations))
          exit(EXIT_FAILURE);
      }
    else
      {
       if(ParseOSMFile(fd,OSMNodes,OSMWays,OSMRelations))
          exit(EXIT_FAILURE);
      }

    CloseFile(fd);

    free(filename);
   }

 DeleteXMLTaggingRules();

 FinishNodeList(OSMNodes);
 FinishWayList(OSMWays);
 FinishRelationList(OSMRelations);

 /* Sort the data */

 printf("\nSort OSM Data\n=============\n\n");
 fflush(stdout);

 /* Sort the nodes, ways and relations */

 SortNodeList(OSMNodes);

 SortWayList(OSMWays);

 SortRelationList(OSMRelations);

 /* Process the data */

 RenameFile(OSMNodes->filename_tmp,OSMNodes->filename);
 RenameFile(OSMWays->filename_tmp,OSMWays->filename);
 RenameFile(OSMRelations->rrfilename_tmp,OSMRelations->rrfilename);
 RenameFile(OSMRelations->trfilename_tmp,OSMRelations->trfilename);

 close_errorlog();

 printf("\nCreate Error Log\n================\n\n");
 fflush(stdout);

 OSMErrorLogs=NewErrorLogList();

 ProcessErrorLogs(OSMErrorLogs,OSMNodes,OSMWays,OSMRelations);

 SortErrorLogsGeographically(OSMErrorLogs);

 SaveErrorLogs(OSMErrorLogs,FileName(dirname,prefix,"fixme.mem"));

 FreeErrorLogList(OSMErrorLogs);

 /* Free the memory (delete the temporary files) */

 FreeNodeList(OSMNodes,0);
 FreeWayList(OSMWays,0);
 FreeRelationList(OSMRelations,0);

 /* Print the total time */

 if(option_logtime)
   {
    printf("\n");
    fprintf_elapsed_time(stdout,&start_time);
    printf("Complete\n");
    fflush(stdout);
   }

 return(0);
}


/*++++++++++++++++++++++++++++++++++++++
  Print out the usage information.

  int detail The level of detail to use - 0 = low, 1 = high.

  const char *argerr The argument that gave the error (if there is one).

  const char *err Other error message (if there is one).
  ++++++++++++++++++++++++++++++++++++++*/

static void print_usage(int detail,const char *argerr,const char *err)
{
 fprintf(stderr,
         "Usage: fixme-finder [--help]\n"
         "                    [--dir=<dirname>]\n"
#if defined(USE_PTHREADS) && USE_PTHREADS
         "                    [--sort-ram-size=<size>] [--sort-threads=<number>]\n"
#else
         "                    [--sort-ram-size=<size>]\n"
#endif
         "                    [--tmpdir=<dirname>]\n"
         "                    [--tagging=<filename>]\n"
         "                    [--loggable] [--logtime]\n"
         "                    [<filename.osm> ...\n"
         "                     | <filename.pbf> ...\n"
         "                     | <filename.o5m> ..."
#if defined(USE_BZIP2) && USE_BZIP2
         "\n                     | <filename.(osm|o5m).bz2> ..."
#endif
#if defined(USE_GZIP) && USE_GZIP
         "\n                     | <filename.(osm|o5m).gz> ..."
#endif
         "]\n");

 if(argerr)
    fprintf(stderr,
            "\n"
            "Error with command line parameter: %s\n",argerr);

 if(err)
    fprintf(stderr,
            "\n"
            "Error: %s\n",err);

 if(detail)
    fprintf(stderr,
            "\n"
            "--help                    Prints this information.\n"
            "\n"
            "--dir=<dirname>           The directory containing the fixme database.\n"
            "\n"
            "--sort-ram-size=<size>    The amount of RAM (in MB) to use for data sorting\n"
#if SLIM
            "                          (defaults to 64MB otherwise.)\n"
#else
            "                          (defaults to 256MB otherwise.)\n"
#endif
#if defined(USE_PTHREADS) && USE_PTHREADS
            "--sort-threads=<number>   The number of threads to use for data sorting.\n"
#endif
            "\n"
            "--tmpdir=<dirname>        The directory name for temporary files.\n"
            "                          (defaults to the '--dir' option directory.)\n"
            "\n"
            "--tagging=<filename>      The name of the XML file containing the tagging rules\n"
            "                          (defaults to 'fixme.xml' with '--dir' option)\n"
            "\n"
            "--loggable                Print progress messages suitable for logging to file.\n"
            "--logtime                 Print the elapsed time for each processing step.\n"
            "\n"
            "<filename.osm>, <filename.pbf>, <filename.o5m>\n"
            "                          The name(s) of the file(s) to read and parse.\n"
            "                          Filenames ending '.pbf' read as PBF, filenames ending\n"
            "                          '.o5m' read as O5M, others as XML.\n"
#if defined(USE_BZIP2) && USE_BZIP2
            "                          Filenames ending '.bz2' will be bzip2 uncompressed.\n"
#endif
#if defined(USE_GZIP) && USE_GZIP
            "                          Filenames ending '.gz' will be gzip uncompressed.\n"
#endif
            );

 exit(!detail);
}
1	/***************************************
2	OSM planet file fixme finder.
3
4	Part of the Routino routing software.
5	****************/ /****************
6	This file Copyright 2008-2013 Andrew M. Bishop
7
8	This program is free software: you can redistribute it and/or modify
9	it under the terms of the GNU Affero General Public License as published by
10	the Free Software Foundation, either version 3 of the License, or
11	(at your option) any later version.
12
13	This program is distributed in the hope that it will be useful,
14	but WITHOUT ANY WARRANTY; without even the implied warranty of
15	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16	GNU Affero General Public License for more details.
17
18	You should have received a copy of the GNU Affero General Public License
19	along with this program. If not, see <http://www.gnu.org/licenses/>.
20	***************************************/
21
22
23	#include <stdio.h>
24	#include <stdlib.h>
25	#include <unistd.h>
26	#include <string.h>
27	#include <errno.h>
28	#include <sys/time.h>
29
30	#include "types.h"
31	#include "ways.h"
32
33	#include "typesx.h"
34	#include "nodesx.h"
35	#include "waysx.h"
36	#include "relationsx.h"
37
38	#include "files.h"
39	#include "logging.h"
40	#include "errorlogx.h"
41	#include "functions.h"
42	#include "osmparser.h"
43	#include "tagging.h"
44	#include "uncompress.h"
45
46
47	/* Global variables */
48
49	/+ The name of the temporary directory. +/
50	char *option_tmpdirname=NULL;
51
52	/+ The amount of RAM to use for filesorting. +/
53	size_t option_filesort_ramsize=0;
54
55	/+ The number of threads to use for filesorting. +/
56	int option_filesort_threads=1;
57
58
59	/* Local functions */
60
61	static void print_usage(int detail,const char argerr,const char err);
62
63
64	/*++++++++++++++++++++++++++++++++++++++
65	The main program for the find-fixme.
66	++++++++++++++++++++++++++++++++++++++*/
67
68	int main(int argc,char** argv)
69	{
70	struct timeval start_time;
71	NodesX *OSMNodes;
72	WaysX *OSMWays;
73	RelationsX *OSMRelations;
74	ErrorLogsX *OSMErrorLogs;
75	char dirname=NULL,prefix=NULL,tagging="fixme.xml",errorlog="fixme.log";
76	int option_keep=1;
77	int option_filenames=0;
78	int arg;
79
80	gettimeofday(&start_time,NULL);
81
82	/* Parse the command line arguments */
83
84	for(arg=1;arg<argc;arg++)
85	{
86	if(!strcmp(argv[arg],"--help"))
87	print_usage(1,NULL,NULL);
88	else if(!strncmp(argv[arg],"--dir=",6))
89	dirname=&argv[arg][6];
90	else if(!strncmp(argv[arg],"--sort-ram-size=",16))
91	option_filesort_ramsize=atoi(&argv[arg][16]);
92	#if defined(USE_PTHREADS) && USE_PTHREADS
93	else if(!strncmp(argv[arg],"--sort-threads=",15))
94	option_filesort_threads=atoi(&argv[arg][15]);
95	#endif
96	else if(!strncmp(argv[arg],"--tmpdir=",9))
97	option_tmpdirname=&argv[arg][9];
98	else if(!strncmp(argv[arg],"--tagging=",10))
99	tagging=&argv[arg][10];
100	else if(!strcmp(argv[arg],"--loggable"))
101	option_loggable=1;
102	else if(!strcmp(argv[arg],"--logtime"))
103	option_logtime=1;
104	else if(argv[arg][0]=='-' && argv[arg][1]=='-')
105	print_usage(0,argv[arg],NULL);
106	else
107	option_filenames++;
108	}
109
110	/* Check the specified command line options */
111
112	if(!option_filesort_ramsize)
113	{
114	#if SLIM
115	option_filesort_ramsize=6410241024;
116	#else
117	option_filesort_ramsize=25610241024;
118	#endif
119	}
120	else
121	option_filesort_ramsize=10241024;
122
123	if(!option_tmpdirname)
124	{
125	if(!dirname)
126	option_tmpdirname=".";
127	else
128	option_tmpdirname=dirname;
129	}
130
131	if(tagging)
132	{
133	if(!ExistsFile(tagging))
134	{
135	fprintf(stderr,"Error: The '--tagging' option specifies a file that does not exist.\n");
136	exit(EXIT_FAILURE);
137	}
138	}
139	else
140	{
141	if(ExistsFile(FileName(dirname,prefix,"tagging.xml")))
142	tagging=FileName(dirname,prefix,"tagging.xml");
143	else
144	{
145	fprintf(stderr,"Error: The '--tagging' option was not used and the default 'tagging.xml' does not exist.\n");
146	exit(EXIT_FAILURE);
147	}
148	}
149
150	if(ParseXMLTaggingRules(tagging))
151	{
152	fprintf(stderr,"Error: Cannot read the tagging rules in the file '%s'.\n",tagging);
153	exit(EXIT_FAILURE);
154	}
155
156	/* Create new node, segment, way and relation variables */
157
158	OSMNodes=NewNodeList(0,0);
159
160	OSMWays=NewWayList(0,0);
161
162	OSMRelations=NewRelationList(0,0);
163
164	/* Create the error log file */
165
166	if(errorlog)
167	open_errorlog(FileName(dirname,prefix,errorlog),0,option_keep);
168
169	/* Parse the file */
170
171	for(arg=1;arg<argc;arg++)
172	{
173	int fd;
174	char filename,p;
175
176	if(argv[arg][0]=='-' && argv[arg][1]=='-')
177	continue;
178
179	filename=strcpy(malloc(strlen(argv[arg])+1),argv[arg]);
180
181	fd=ReOpenFile(filename);
182
183	if((p=strstr(filename,".bz2")) && !strcmp(p,".bz2"))
184	{
185	fd=Uncompress_Bzip2(fd);
186	*p=0;
187	}
188
189	if((p=strstr(filename,".gz")) && !strcmp(p,".gz"))
190	{
191	fd=Uncompress_Gzip(fd);
192	*p=0;
193	}
194
195	printf("\nParse OSM Data [%s]\n==============\n\n",filename);
196	fflush(stdout);
197
198	if((p=strstr(filename,".pbf")) && !strcmp(p,".pbf"))
199	{
200	if(ParsePBFFile(fd,OSMNodes,OSMWays,OSMRelations))
201	exit(EXIT_FAILURE);
202	}
203	else if((p=strstr(filename,".o5m")) && !strcmp(p,".o5m"))
204	{
205	if(ParseO5MFile(fd,OSMNodes,OSMWays,OSMRelations))
206	exit(EXIT_FAILURE);
207	}
208	else
209	{
210	if(ParseOSMFile(fd,OSMNodes,OSMWays,OSMRelations))
211	exit(EXIT_FAILURE);
212	}
213
214	CloseFile(fd);
215
216	free(filename);
217	}
218
219	DeleteXMLTaggingRules();
220
221	FinishNodeList(OSMNodes);
222	FinishWayList(OSMWays);
223	FinishRelationList(OSMRelations);
224
225	/* Sort the data */
226
227	printf("\nSort OSM Data\n=============\n\n");
228	fflush(stdout);
229
230	/* Sort the nodes, ways and relations */
231
232	SortNodeList(OSMNodes);
233
234	SortWayList(OSMWays);
235
236	SortRelationList(OSMRelations);
237
238	/* Process the data */
239
240	RenameFile(OSMNodes->filename_tmp,OSMNodes->filename);
241	RenameFile(OSMWays->filename_tmp,OSMWays->filename);
242	RenameFile(OSMRelations->rrfilename_tmp,OSMRelations->rrfilename);
243	RenameFile(OSMRelations->trfilename_tmp,OSMRelations->trfilename);
244
245	close_errorlog();
246
247	printf("\nCreate Error Log\n================\n\n");
248	fflush(stdout);
249
250	OSMErrorLogs=NewErrorLogList();
251
252	ProcessErrorLogs(OSMErrorLogs,OSMNodes,OSMWays,OSMRelations);
253
254	SortErrorLogsGeographically(OSMErrorLogs);
255
256	SaveErrorLogs(OSMErrorLogs,FileName(dirname,prefix,"fixme.mem"));
257
258	FreeErrorLogList(OSMErrorLogs);
259
260	/* Free the memory (delete the temporary files) */
261
262	FreeNodeList(OSMNodes,0);
263	FreeWayList(OSMWays,0);
264	FreeRelationList(OSMRelations,0);
265
266	/* Print the total time */
267
268	if(option_logtime)
269	{
270	printf("\n");
271	fprintf_elapsed_time(stdout,&start_time);
272	printf("Complete\n");
273	fflush(stdout);
274	}
275
276	return(0);
277	}
278
279
280	/*++++++++++++++++++++++++++++++++++++++
281	Print out the usage information.
282
283	int detail The level of detail to use - 0 = low, 1 = high.
284
285	const char *argerr The argument that gave the error (if there is one).
286
287	const char *err Other error message (if there is one).
288	++++++++++++++++++++++++++++++++++++++*/
289
290	static void print_usage(int detail,const char argerr,const char err)
291	{
292	fprintf(stderr,
293	"Usage: fixme-finder [--help]\n"
294	" [--dir=<dirname>]\n"
295	#if defined(USE_PTHREADS) && USE_PTHREADS
296	" [--sort-ram-size=<size>] [--sort-threads=<number>]\n"
297	#else
298	" [--sort-ram-size=<size>]\n"
299	#endif
300	" [--tmpdir=<dirname>]\n"
301	" [--tagging=<filename>]\n"
302	" [--loggable] [--logtime]\n"
303	" [<filename.osm> ...\n"
304	" \| <filename.pbf> ...\n"
305	" \| <filename.o5m> ..."
306	#if defined(USE_BZIP2) && USE_BZIP2
307	"\n \| <filename.(osm\|o5m).bz2> ..."
308	#endif
309	#if defined(USE_GZIP) && USE_GZIP
310	"\n \| <filename.(osm\|o5m).gz> ..."
311	#endif
312	"]\n");
313
314	if(argerr)
315	fprintf(stderr,
316	"\n"
317	"Error with command line parameter: %s\n",argerr);
318
319	if(err)
320	fprintf(stderr,
321	"\n"
322	"Error: %s\n",err);
323
324	if(detail)
325	fprintf(stderr,
326	"\n"
327	"--help Prints this information.\n"
328	"\n"
329	"--dir=<dirname> The directory containing the fixme database.\n"
330	"\n"
331	"--sort-ram-size=<size> The amount of RAM (in MB) to use for data sorting\n"
332	#if SLIM
333	" (defaults to 64MB otherwise.)\n"
334	#else
335	" (defaults to 256MB otherwise.)\n"
336	#endif
337	#if defined(USE_PTHREADS) && USE_PTHREADS
338	"--sort-threads=<number> The number of threads to use for data sorting.\n"
339	#endif
340	"\n"
341	"--tmpdir=<dirname> The directory name for temporary files.\n"
342	" (defaults to the '--dir' option directory.)\n"
343	"\n"
344	"--tagging=<filename> The name of the XML file containing the tagging rules\n"
345	" (defaults to 'fixme.xml' with '--dir' option)\n"
346	"\n"
347	"--loggable Print progress messages suitable for logging to file.\n"
348	"--logtime Print the elapsed time for each processing step.\n"
349	"\n"
350	"<filename.osm>, <filename.pbf>, <filename.o5m>\n"
351	" The name(s) of the file(s) to read and parse.\n"
352	" Filenames ending '.pbf' read as PBF, filenames ending\n"
353	" '.o5m' read as O5M, others as XML.\n"
354	#if defined(USE_BZIP2) && USE_BZIP2
355	" Filenames ending '.bz2' will be bzip2 uncompressed.\n"
356	#endif
357	#if defined(USE_GZIP) && USE_GZIP
358	" Filenames ending '.gz' will be gzip uncompressed.\n"
359	#endif
360	);
361
362	exit(!detail);
363	}