extras/find-fixme/fixme-finder.c

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

 Part of the Routino routing software.
 ******************/ /******************
 This file Copyright 2008-2014 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 "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)
{
 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;

 printf_program_start();

 /* 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(!strcmp(argv[arg],"--logmemory"))
       option_logmemory=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=OpenFile(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;
      }

    if((p=strstr(filename,".xz")) && !strcmp(p,".xz"))
      {
       fd=Uncompress_Xz(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);

 printf_program_end();

 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] [--logmemory]\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
#if defined(USE_XZ) && USE_XZ
         "\n                     | <filename.(osm|o5m).xz> ..."
#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"
            "--logmemory               Print the max allocated/mapped memory for each 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
#if defined(USE_XZ) && USE_XZ
            "                          Filenames ending '.xz' will be xz 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 2008-2014 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
29	#include "types.h"
30	#include "ways.h"
31
32	#include "typesx.h"
33	#include "nodesx.h"
34	#include "waysx.h"
35	#include "relationsx.h"
36
37	#include "files.h"
38	#include "logging.h"
39	#include "errorlogx.h"
40	#include "functions.h"
41	#include "osmparser.h"
42	#include "tagging.h"
43	#include "uncompress.h"
44
45
46	/* Global variables */
47
48	/+ The name of the temporary directory. +/
49	char *option_tmpdirname=NULL;
50
51	/+ The amount of RAM to use for filesorting. +/
52	size_t option_filesort_ramsize=0;
53
54	/+ The number of threads to use for filesorting. +/
55	int option_filesort_threads=1;
56
57
58	/* Local functions */
59
60	static void print_usage(int detail,const char argerr,const char err);
61
62
63	/*++++++++++++++++++++++++++++++++++++++
64	The main program for the find-fixme.
65	++++++++++++++++++++++++++++++++++++++*/
66
67	int main(int argc,char** argv)
68	{
69	NodesX *OSMNodes;
70	WaysX *OSMWays;
71	RelationsX *OSMRelations;
72	ErrorLogsX *OSMErrorLogs;
73	char dirname=NULL,prefix=NULL,tagging="fixme.xml",errorlog="fixme.log";
74	int option_keep=1;
75	int option_filenames=0;
76	int arg;
77
78	printf_program_start();
79
80	/* Parse the command line arguments */
81
82	for(arg=1;arg<argc;arg++)
83	{
84	if(!strcmp(argv[arg],"--help"))
85	print_usage(1,NULL,NULL);
86	else if(!strncmp(argv[arg],"--dir=",6))
87	dirname=&argv[arg][6];
88	else if(!strncmp(argv[arg],"--sort-ram-size=",16))
89	option_filesort_ramsize=atoi(&argv[arg][16]);
90	#if defined(USE_PTHREADS) && USE_PTHREADS
91	else if(!strncmp(argv[arg],"--sort-threads=",15))
92	option_filesort_threads=atoi(&argv[arg][15]);
93	#endif
94	else if(!strncmp(argv[arg],"--tmpdir=",9))
95	option_tmpdirname=&argv[arg][9];
96	else if(!strncmp(argv[arg],"--tagging=",10))
97	tagging=&argv[arg][10];
98	else if(!strcmp(argv[arg],"--loggable"))
99	option_loggable=1;
100	else if(!strcmp(argv[arg],"--logtime"))
101	option_logtime=1;
102	else if(!strcmp(argv[arg],"--logmemory"))
103	option_logmemory=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=OpenFile(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	if((p=strstr(filename,".xz")) && !strcmp(p,".xz"))
196	{
197	fd=Uncompress_Xz(fd);
198	*p=0;
199	}
200
201	printf("\nParse OSM Data [%s]\n==============\n\n",filename);
202	fflush(stdout);
203
204	if((p=strstr(filename,".pbf")) && !strcmp(p,".pbf"))
205	{
206	if(ParsePBFFile(fd,OSMNodes,OSMWays,OSMRelations))
207	exit(EXIT_FAILURE);
208	}
209	else if((p=strstr(filename,".o5m")) && !strcmp(p,".o5m"))
210	{
211	if(ParseO5MFile(fd,OSMNodes,OSMWays,OSMRelations))
212	exit(EXIT_FAILURE);
213	}
214	else
215	{
216	if(ParseOSMFile(fd,OSMNodes,OSMWays,OSMRelations))
217	exit(EXIT_FAILURE);
218	}
219
220	CloseFile(fd);
221
222	free(filename);
223	}
224
225	DeleteXMLTaggingRules();
226
227	FinishNodeList(OSMNodes);
228	FinishWayList(OSMWays);
229	FinishRelationList(OSMRelations);
230
231	/* Sort the data */
232
233	printf("\nSort OSM Data\n=============\n\n");
234	fflush(stdout);
235
236	/* Sort the nodes, ways and relations */
237
238	SortNodeList(OSMNodes);
239
240	SortWayList(OSMWays);
241
242	SortRelationList(OSMRelations);
243
244	/* Process the data */
245
246	RenameFile(OSMNodes->filename_tmp,OSMNodes->filename);
247	RenameFile(OSMWays->filename_tmp,OSMWays->filename);
248	RenameFile(OSMRelations->rrfilename_tmp,OSMRelations->rrfilename);
249	RenameFile(OSMRelations->trfilename_tmp,OSMRelations->trfilename);
250
251	close_errorlog();
252
253	printf("\nCreate Error Log\n================\n\n");
254	fflush(stdout);
255
256	OSMErrorLogs=NewErrorLogList();
257
258	ProcessErrorLogs(OSMErrorLogs,OSMNodes,OSMWays,OSMRelations);
259
260	SortErrorLogsGeographically(OSMErrorLogs);
261
262	SaveErrorLogs(OSMErrorLogs,FileName(dirname,prefix,"fixme.mem"));
263
264	FreeErrorLogList(OSMErrorLogs);
265
266	/* Free the memory (delete the temporary files) */
267
268	FreeNodeList(OSMNodes,0);
269	FreeWayList(OSMWays,0);
270	FreeRelationList(OSMRelations,0);
271
272	printf_program_end();
273
274	return(0);
275	}
276
277
278	/*++++++++++++++++++++++++++++++++++++++
279	Print out the usage information.
280
281	int detail The level of detail to use - 0 = low, 1 = high.
282
283	const char *argerr The argument that gave the error (if there is one).
284
285	const char *err Other error message (if there is one).
286	++++++++++++++++++++++++++++++++++++++*/
287
288	static void print_usage(int detail,const char argerr,const char err)
289	{
290	fprintf(stderr,
291	"Usage: fixme-finder [--help]\n"
292	" [--dir=<dirname>]\n"
293	#if defined(USE_PTHREADS) && USE_PTHREADS
294	" [--sort-ram-size=<size>] [--sort-threads=<number>]\n"
295	#else
296	" [--sort-ram-size=<size>]\n"
297	#endif
298	" [--tmpdir=<dirname>]\n"
299	" [--tagging=<filename>]\n"
300	" [--loggable] [--logtime] [--logmemory]\n"
301	" [<filename.osm> ...\n"
302	" \| <filename.pbf> ...\n"
303	" \| <filename.o5m> ..."
304	#if defined(USE_BZIP2) && USE_BZIP2
305	"\n \| <filename.(osm\|o5m).bz2> ..."
306	#endif
307	#if defined(USE_GZIP) && USE_GZIP
308	"\n \| <filename.(osm\|o5m).gz> ..."
309	#endif
310	#if defined(USE_XZ) && USE_XZ
311	"\n \| <filename.(osm\|o5m).xz> ..."
312	#endif
313	"]\n");
314
315	if(argerr)
316	fprintf(stderr,
317	"\n"
318	"Error with command line parameter: %s\n",argerr);
319
320	if(err)
321	fprintf(stderr,
322	"\n"
323	"Error: %s\n",err);
324
325	if(detail)
326	fprintf(stderr,
327	"\n"
328	"--help Prints this information.\n"
329	"\n"
330	"--dir=<dirname> The directory containing the fixme database.\n"
331	"\n"
332	"--sort-ram-size=<size> The amount of RAM (in MB) to use for data sorting\n"
333	#if SLIM
334	" (defaults to 64MB otherwise.)\n"
335	#else
336	" (defaults to 256MB otherwise.)\n"
337	#endif
338	#if defined(USE_PTHREADS) && USE_PTHREADS
339	"--sort-threads=<number> The number of threads to use for data sorting.\n"
340	#endif
341	"\n"
342	"--tmpdir=<dirname> The directory name for temporary files.\n"
343	" (defaults to the '--dir' option directory.)\n"
344	"\n"
345	"--tagging=<filename> The name of the XML file containing the tagging rules\n"
346	" (defaults to 'fixme.xml' with '--dir' option)\n"
347	"\n"
348	"--loggable Print progress messages suitable for logging to file.\n"
349	"--logtime Print the elapsed time for each processing step.\n"
350	"--logmemory Print the max allocated/mapped memory for each step.\n"
351	"\n"
352	"<filename.osm>, <filename.pbf>, <filename.o5m>\n"
353	" The name(s) of the file(s) to read and parse.\n"
354	" Filenames ending '.pbf' read as PBF, filenames ending\n"
355	" '.o5m' read as O5M, others as XML.\n"
356	#if defined(USE_BZIP2) && USE_BZIP2
357	" Filenames ending '.bz2' will be bzip2 uncompressed.\n"
358	#endif
359	#if defined(USE_GZIP) && USE_GZIP
360	" Filenames ending '.gz' will be gzip uncompressed.\n"
361	#endif
362	#if defined(USE_XZ) && USE_XZ
363	" Filenames ending '.xz' will be xz uncompressed.\n"
364	#endif
365	);
366
367	exit(!detail);
368	}