extras/find-fixme/fixme-finder.c

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

 Part of the Routino routing software.
 ******************/ /******************
 This file Copyright 2008-2015 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 <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=NULL;
 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
   {
    tagging=FileName(dirname,prefix,"fixme.xml");

    if(!ExistsFile(tagging))
      {
       fprintf(stderr,"Error: The '--tagging' option was not used and the default 'fixme.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 */

 open_errorlog(FileName(dirname,prefix,"fixme.log"),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("\n");
 fflush(stdout);

 printf_program_end();

 exit(EXIT_SUCCESS);
}


/*++++++++++++++++++++++++++++++++++++++
  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 2008-2015 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 <string.h>
26	#include <errno.h>
27
28	#include "types.h"
29	#include "ways.h"
30
31	#include "typesx.h"
32	#include "nodesx.h"
33	#include "waysx.h"
34	#include "relationsx.h"
35
36	#include "files.h"
37	#include "logging.h"
38	#include "errorlogx.h"
39	#include "functions.h"
40	#include "osmparser.h"
41	#include "tagging.h"
42	#include "uncompress.h"
43
44
45	/* Global variables */
46
47	/+ The name of the temporary directory. +/
48	char *option_tmpdirname=NULL;
49
50	/+ The amount of RAM to use for filesorting. +/
51	size_t option_filesort_ramsize=0;
52
53	/+ The number of threads to use for filesorting. +/
54	int option_filesort_threads=1;
55
56
57	/* Local functions */
58
59	static void print_usage(int detail,const char argerr,const char err);
60
61
62	/*++++++++++++++++++++++++++++++++++++++
63	The main program for the find-fixme.
64	++++++++++++++++++++++++++++++++++++++*/
65
66	int main(int argc,char** argv)
67	{
68	NodesX *OSMNodes;
69	WaysX *OSMWays;
70	RelationsX *OSMRelations;
71	ErrorLogsX *OSMErrorLogs;
72	char dirname=NULL,prefix=NULL,*tagging=NULL;
73	int option_keep=1;
74	int option_filenames=0;
75	int arg;
76
77	printf_program_start();
78
79	/* Parse the command line arguments */
80
81	for(arg=1;arg<argc;arg++)
82	{
83	if(!strcmp(argv[arg],"--help"))
84	print_usage(1,NULL,NULL);
85	else if(!strncmp(argv[arg],"--dir=",6))
86	dirname=&argv[arg][6];
87	else if(!strncmp(argv[arg],"--sort-ram-size=",16))
88	option_filesort_ramsize=atoi(&argv[arg][16]);
89	#if defined(USE_PTHREADS) && USE_PTHREADS
90	else if(!strncmp(argv[arg],"--sort-threads=",15))
91	option_filesort_threads=atoi(&argv[arg][15]);
92	#endif
93	else if(!strncmp(argv[arg],"--tmpdir=",9))
94	option_tmpdirname=&argv[arg][9];
95	else if(!strncmp(argv[arg],"--tagging=",10))
96	tagging=&argv[arg][10];
97	else if(!strcmp(argv[arg],"--loggable"))
98	option_loggable=1;
99	else if(!strcmp(argv[arg],"--logtime"))
100	option_logtime=1;
101	else if(!strcmp(argv[arg],"--logmemory"))
102	option_logmemory=1;
103	else if(argv[arg][0]=='-' && argv[arg][1]=='-')
104	print_usage(0,argv[arg],NULL);
105	else
106	option_filenames++;
107	}
108
109	/* Check the specified command line options */
110
111	if(!option_filesort_ramsize)
112	{
113	#if SLIM
114	option_filesort_ramsize=6410241024;
115	#else
116	option_filesort_ramsize=25610241024;
117	#endif
118	}
119	else
120	option_filesort_ramsize=10241024;
121
122	if(!option_tmpdirname)
123	{
124	if(!dirname)
125	option_tmpdirname=".";
126	else
127	option_tmpdirname=dirname;
128	}
129
130	if(tagging)
131	{
132	if(!ExistsFile(tagging))
133	{
134	fprintf(stderr,"Error: The '--tagging' option specifies a file that does not exist.\n");
135	exit(EXIT_FAILURE);
136	}
137	}
138	else
139	{
140	tagging=FileName(dirname,prefix,"fixme.xml");
141
142	if(!ExistsFile(tagging))
143	{
144	fprintf(stderr,"Error: The '--tagging' option was not used and the default 'fixme.xml' does not exist.\n");
145	exit(EXIT_FAILURE);
146	}
147	}
148
149	if(ParseXMLTaggingRules(tagging))
150	{
151	fprintf(stderr,"Error: Cannot read the tagging rules in the file '%s'.\n",tagging);
152	exit(EXIT_FAILURE);
153	}
154
155	/* Create new node, segment, way and relation variables */
156
157	OSMNodes=NewNodeList(0,0);
158
159	OSMWays=NewWayList(0,0);
160
161	OSMRelations=NewRelationList(0,0);
162
163	/* Create the error log file */
164
165	open_errorlog(FileName(dirname,prefix,"fixme.log"),0,option_keep);
166
167	/* Parse the file */
168
169	for(arg=1;arg<argc;arg++)
170	{
171	int fd;
172	char filename,p;
173
174	if(argv[arg][0]=='-' && argv[arg][1]=='-')
175	continue;
176
177	filename=strcpy(malloc(strlen(argv[arg])+1),argv[arg]);
178
179	fd=OpenFile(filename);
180
181	if((p=strstr(filename,".bz2")) && !strcmp(p,".bz2"))
182	{
183	fd=Uncompress_Bzip2(fd);
184	*p=0;
185	}
186
187	if((p=strstr(filename,".gz")) && !strcmp(p,".gz"))
188	{
189	fd=Uncompress_Gzip(fd);
190	*p=0;
191	}
192
193	if((p=strstr(filename,".xz")) && !strcmp(p,".xz"))
194	{
195	fd=Uncompress_Xz(fd);
196	*p=0;
197	}
198
199	printf("\nParse OSM Data [%s]\n==============\n\n",filename);
200	fflush(stdout);
201
202	if((p=strstr(filename,".pbf")) && !strcmp(p,".pbf"))
203	{
204	if(ParsePBFFile(fd,OSMNodes,OSMWays,OSMRelations))
205	exit(EXIT_FAILURE);
206	}
207	else if((p=strstr(filename,".o5m")) && !strcmp(p,".o5m"))
208	{
209	if(ParseO5MFile(fd,OSMNodes,OSMWays,OSMRelations))
210	exit(EXIT_FAILURE);
211	}
212	else
213	{
214	if(ParseOSMFile(fd,OSMNodes,OSMWays,OSMRelations))
215	exit(EXIT_FAILURE);
216	}
217
218	CloseFile(fd);
219
220	free(filename);
221	}
222
223	DeleteXMLTaggingRules();
224
225	FinishNodeList(OSMNodes);
226	FinishWayList(OSMWays);
227	FinishRelationList(OSMRelations);
228
229	/* Sort the data */
230
231	printf("\nSort OSM Data\n=============\n\n");
232	fflush(stdout);
233
234	/* Sort the nodes, ways and relations */
235
236	SortNodeList(OSMNodes);
237
238	SortWayList(OSMWays);
239
240	SortRelationList(OSMRelations);
241
242	/* Process the data */
243
244	RenameFile(OSMNodes->filename_tmp,OSMNodes->filename);
245	RenameFile(OSMWays->filename_tmp,OSMWays->filename);
246	RenameFile(OSMRelations->rrfilename_tmp,OSMRelations->rrfilename);
247	RenameFile(OSMRelations->trfilename_tmp,OSMRelations->trfilename);
248
249	close_errorlog();
250
251	printf("\nCreate Error Log\n================\n\n");
252	fflush(stdout);
253
254	OSMErrorLogs=NewErrorLogList();
255
256	ProcessErrorLogs(OSMErrorLogs,OSMNodes,OSMWays,OSMRelations);
257
258	SortErrorLogsGeographically(OSMErrorLogs);
259
260	SaveErrorLogs(OSMErrorLogs,FileName(dirname,prefix,"fixme.mem"));
261
262	FreeErrorLogList(OSMErrorLogs);
263
264	/* Free the memory (delete the temporary files) */
265
266	FreeNodeList(OSMNodes,0);
267	FreeWayList(OSMWays,0);
268	FreeRelationList(OSMRelations,0);
269
270	printf("\n");
271	fflush(stdout);
272
273	printf_program_end();
274
275	exit(EXIT_SUCCESS);
276	}
277
278
279	/*++++++++++++++++++++++++++++++++++++++
280	Print out the usage information.
281
282	int detail The level of detail to use - 0 = low, 1 = high.
283
284	const char *argerr The argument that gave the error (if there is one).
285
286	const char *err Other error message (if there is one).
287	++++++++++++++++++++++++++++++++++++++*/
288
289	static void print_usage(int detail,const char argerr,const char err)
290	{
291	fprintf(stderr,
292	"Usage: fixme-finder [--help]\n"
293	" [--dir=<dirname>]\n"
294	#if defined(USE_PTHREADS) && USE_PTHREADS
295	" [--sort-ram-size=<size>] [--sort-threads=<number>]\n"
296	#else
297	" [--sort-ram-size=<size>]\n"
298	#endif
299	" [--tmpdir=<dirname>]\n"
300	" [--tagging=<filename>]\n"
301	" [--loggable] [--logtime] [--logmemory]\n"
302	" [<filename.osm> ...\n"
303	" \| <filename.pbf> ...\n"
304	" \| <filename.o5m> ..."
305	#if defined(USE_BZIP2) && USE_BZIP2
306	"\n \| <filename.(osm\|o5m).bz2> ..."
307	#endif
308	#if defined(USE_GZIP) && USE_GZIP
309	"\n \| <filename.(osm\|o5m).gz> ..."
310	#endif
311	#if defined(USE_XZ) && USE_XZ
312	"\n \| <filename.(osm\|o5m).xz> ..."
313	#endif
314	"]\n");
315
316	if(argerr)
317	fprintf(stderr,
318	"\n"
319	"Error with command line parameter: %s\n",argerr);
320
321	if(err)
322	fprintf(stderr,
323	"\n"
324	"Error: %s\n",err);
325
326	if(detail)
327	fprintf(stderr,
328	"\n"
329	"--help Prints this information.\n"
330	"\n"
331	"--dir=<dirname> The directory containing the fixme database.\n"
332	"\n"
333	"--sort-ram-size=<size> The amount of RAM (in MB) to use for data sorting\n"
334	#if SLIM
335	" (defaults to 64MB otherwise.)\n"
336	#else
337	" (defaults to 256MB otherwise.)\n"
338	#endif
339	#if defined(USE_PTHREADS) && USE_PTHREADS
340	"--sort-threads=<number> The number of threads to use for data sorting.\n"
341	#endif
342	"\n"
343	"--tmpdir=<dirname> The directory name for temporary files.\n"
344	" (defaults to the '--dir' option directory.)\n"
345	"\n"
346	"--tagging=<filename> The name of the XML file containing the tagging rules\n"
347	" (defaults to 'fixme.xml' with '--dir' option)\n"
348	"\n"
349	"--loggable Print progress messages suitable for logging to file.\n"
350	"--logtime Print the elapsed time for each processing step.\n"
351	"--logmemory Print the max allocated/mapped memory for each step.\n"
352	"\n"
353	"<filename.osm>, <filename.pbf>, <filename.o5m>\n"
354	" The name(s) of the file(s) to read and parse.\n"
355	" Filenames ending '.pbf' read as PBF, filenames ending\n"
356	" '.o5m' read as O5M, others as XML.\n"
357	#if defined(USE_BZIP2) && USE_BZIP2
358	" Filenames ending '.bz2' will be bzip2 uncompressed.\n"
359	#endif
360	#if defined(USE_GZIP) && USE_GZIP
361	" Filenames ending '.gz' will be gzip uncompressed.\n"
362	#endif
363	#if defined(USE_XZ) && USE_XZ
364	" Filenames ending '.xz' will be xz uncompressed.\n"
365	#endif
366	);
367
368	exit(!detail);
369	}