The format used for Garmin maps has, in effect, a maximum size, meaning that you have to split an .osm file that contains large well mapped regions into a number of smaller tiles. This program does that. There are at least two stages of processing required. The first stage is to calculate what area each tile should cover, based on the distribution of nodes. The second stage writes out the nodes, ways and relations from the original .osm file into separate smaller .osm files, one for each area that was calculated in stage one. With option keep-complete=true, two additional stages are used to avoid broken ways and polygons.
The two most important features are:
- Variable sized tiles so that you don't get a large number of tiny files.
- Tiles join exactly with no overlap or gaps.
You will need a lot of memory on your computer if you intend to split a large area. A few options allow to configure how much memory you need. With the default parameters, you need about 4-5 bytes for every node and way. This doesn't sound a lot but there are about 1700 million nodes in the whole planet file and so you cannot process the whole planet in one pass file on a 32 bit machine using this utility as the maximum java heap space is 2G. It is possible with 64 bit java and about 7GB of heap or with multiple passes.
The Europe extract from Cloudmade or Geofabrik can be processed within the 2G limit if you have sufficient memory. With the default options europe is split into about 750 tiles. The Europe extract is about half of the size of the complete planet file.
On the other hand a single country, even a well mapped one such as Germany or the UK, will be possible on a modest machine, even a netbook.
Download from the splitter download directory
The source code is available from subversion: at http://svn.mkgmap.org.uk/splitter/trunk
Splitter requires java 1.8 or higher. Unzip the downloaded package. A package like splitter-r580.zip contains a directory structure with top level name splitter-r580 containing splitter.jar and sub directories lib and doc. This directory structure is needed. cd into the directory that contains splitter.jar and run the following command. If you have less than 2G of memory on your computer you should reduce the -Xmx argument
java -Xmx2000m -jar splitter.jar file.osm > splitter.log
This will produce a number of .osm.pbf files that can be read by mkgmap. There are also other files produced:
The template.args file is a file that can be used with the -c option of mkgmap that will compile all the files. You can use it as is or you can copy it and edit it to include your own options. For example instead of each description being "OSM Map" it could be "NW Scotland" as appropriate.
The areas.list file is the list of bounding boxes that were calculated. If you want you can use this on a subsequent call the splitter using the --split-file option to use exactly the same areas as last time. This might be useful if you produce a map regularly and want to keep the tile areas the same from month to month. It is also useful to avoid the time it takes to regenerate the file each time (currently about a third of the overall time taken to perform the split). Of course if the map grows enough that one of the tiles overflows you will have to re-calculate the areas again.
The areas.poly file contains the bounding polygon of the calculated areas. See option --polygon-file how this can be used.
The densities-out.txt file is written when no split-file is given and contains debugging information only.
You can also use a gzip'ed or bz2'ed compressed .osm file as the input file. Note that this can slow down the splitter considerably (particularly true for bz2) because decompressing the .osm file can take quite a lot of CPU power. If you are likely to be processing a file several times you're probably better off converting the file to one of the binary formats pbf or o5m. The o5m format is faster to read, but requires more space on the disk.
There are a number of options to fine tune things that you might want to try.
- A comma separated list of boundary tag values for relations. This is used in the gen-problem-list phase to filter relations, see also keep-complete. The filter is applied to relations with
type=boundary and type=multipolygon. The default value "use-exclude-list" means that splitter uses a hard-coded exclude list containing the values "administrative", "postal_code", and "political". So, by default, a relation that has a boundary=administrative tag will not be added to the generated problem list. If you overwrite the default value, you have to specify one or more values that are used as an include list, e.g. --boundary-tags=national_park,protected_area. Boundary relations that are not added to the generated problem list may be incomplete if they cross tile borders. The default setting assumes that you use precompiled bounds in mkgmap. If you use option --location-autofill without option --bounds in mkgmap, make sure to use --boundary-tags=administrative,postal_code and add those tags that you use in your style file.
- Deprecated, now does nothing.
- Sets the desciption to be written in to the template.args file.
- The name of a GeoNames file to use for determining tile names. Typically cities15000.zip from geonames
- Use keep-complete=false to disable two additional program phases between the split and the final distribution phase (not recommended). The first phase, called gen-problem-list, detects all ways and relations that are crossing the borders of one or more output files. The second phase, called handle-problem-list, collects the coordinates of these ways and relations and calculates all output files that are crossed or enclosed. The information is passed to the final dist-phase in three temporary files. This avoids broken polygons, but be aware that it requires to read the input files at least two additional times.
Do not specify it with --overlap unless you have a good reason to do so.
- Set the filename for the split files. In the example the first file will be called 63240001.osm.pbf and the next one will be 63240002.osm.pbf and so on.
- The maximum number of areas that can be processed in a single pass during the second stage of processing. This must be a number from 1 to 9999. Higher numbers mean fewer passes over the source file and hence quicker overall processing, but also require more memory. If you find you are running out of memory but don't want to increase your --max-nodes value, try reducing this instead. Changing this will have no effect on the result of the split, it's purely to let you trade off memory for performance. Note that the first stage of the processing has a fixed memory overhead regardless of what this is set to so if you are running out of memory before the areas.list file is generated, you need to either increase your -Xmx value or reduce the size of the input file you're trying to split.
- The maximum number of nodes that can be in any of the resultant files. The default is fairly conservative, I think you could increase it quite a lot before getting any 'map too big' messages. I've not experimented much. Also the bigger this value, the less memory is required during the splitting stage.
- The maximum number of threads used by splitter. Default is auto.
- Specify this if the input osm file has nodes, ways and relations intermingled or the ids are not strictly sorted. To increase performance, use the osmosis sort function.
- Don't trim empty space off the edges of tiles. This option is ignored when --polygon-file is used.
- A target value that is used when no split-file is given. Splitting is done so that the given number of tiles is produced. The max-nodes value is ignored if this option is given. Added with r323.
- The format in which the output files are written. Possible values are xml, pbf, o5m, and simulate. The default is pbf, which produces the smallest file sizes. The o5m format is faster to write, but creates around 40% larger files. The simulate option is for debugging purposes. Note that the different formats also require different amounts of java heap. The o5m format is worst with ~1.2 MB for each tile, pbf needs~0.5 MB and xml is around 0.4MB.
- The directory to which splitter should write the output files. If the specified path to a directory doesn't exist, splitter tries to create it. Defaults to the current working directory.
- Deprecated since r279. With keep-complete=false, splitter should include nodes outside the bounding box, so that mkgmap can neatly crop exactly at the border. This parameter controls the size of that overlap. It is in map units, a default of 2000 is used which means about 0.04 degrees of latitude or longitude. If --keep-complete=true is active and --overlap is given, a warning will be printed because this combination rarely makes sense.
- The name of a file containing a bounding polygon in osmosis polygon file format. Splitter uses this file when calculating the areas. It first calculates a grid using the given --resolution. The input file is read and for each node, a counter is increased for the related grid area. If the input file contains a bounding box, this is applied to the grid so that nodes outside of the bounding box are ignored. Next, if specified, the bounding polygon is used to zero those grid elements outside of the bounding polygon area. If the polygon area(s) describe(s) a rectilinear area with no more than 40 vertices, splitter will try to create output files that fit exactly into the area, else it will approximate the polygon area with rectangles.
- The name of a directory or zip file containing precompiled sea tiles. If given, splitter will use the precompiled sea tiles in the same way as mkgmap does. Use this if you want to use a polygon-file or --no-trim=true and mkgmap creates empty *.img files combined with a message starting "There is not enough room in a single garmin map for all the input data".
- The name of a file containing ways and relations that are known to cause problems in the split process. Use this option if --keep-complete requires too much time or memory and --overlap doesn't solve your problem.
- Syntax of problem file:
way:<id> # foo rel:<id> # bar
way:2784765 # Ferry Guernsey - Jersey
- The name of a file to write the generated problem list created with --keep-complete. The parameter is ignored if --keep-complete=false. You can reuse this file with the --problem-file parameter, but do this only if you use the same values for max-nodes and resolution.
- The resolution of the density map produced during the first phase. A value between 1 and 24. Default is 13. Increasing the value to 14 requires four times more memory in the split phase. The value is ignored if a --split-file is given.
- Use the previously calculated tile areas instead of calculating them from scratch. The file can also be in *.kml format.
- Displays the amount of memory used by the JVM every --status-freq seconds. Set =0 to disable. Default is 120.
- Debugging: stop after a given program phase. Can be split, gen-problem-list, or handle-problem-list Default is dist which means execute all phases.
- The name of a kml file to write out the areas to. This is in addition to areas.list (which is always written out).
- If the parameter --version is found somewhere in the options, splitter will just print the version info and exit. Version info looks like this:
splitter 279 compiled 2013-01-12T01:45:02+0000
- If the parameter --help is found somewhere in the options, splitter will print a list of all known normal options together with a short help and exit.
Tuning for best performance
A few hints for those that are using splitter to split large files.
- For faster processing with --keep-complete=true, convert the input file to o5m format using
osmconvert --drop-version file.osm -o=file.o5m
- The option --drop-version is optional, it reduces the file to that data that is needed by splitter and mkgmap.
- If you still experience poor performance, look into splitter.log. Search for the word Distributing . You may find something like this in the next line:
Processing 1502 areas in 3 passes, 501 areas at a time
- This means splitter has to read the input file input three times because the max-areas parameter was much smaller than the number of areas. If you have enough heap, set max-areas value to a value that is higher than the number of areas, e.g. --max-areas=2048. Execute splitter again and you should find
Processing 1502 areas in a single pass
- More areas require more memory. Make sure that splitter has enough heap (increase the -Xmx parameter) so that it doesn't waste much time in the garbage collector (GC), but keep as much memory as possible for the systems I/O caches.
- If available, use two different disks for input file and output directory, esp. when you use o5m format for input and output.
- If you use mkgmap r2415 or later and disk space is no concern, consider to use --output=o5m to speed up processing. Note that this requires more memory than pbf or xml.
Tuning for low memory requirements
If your machine has less than 1GB free memory (eg. a netbook), you can still use splitter, but you might have to be patient if you use parameter --keep-complete and want to split a file like germany.osm.pbf or a larger one. If needed, reduce the number of parallel processed areas to 50 with the max-areas parameter. You have to use --keep-complete=false when splitting an area like Europe.
- There is no longer an upper limit on the number of areas that can be output (previously it was 255). More areas just mean potentially more passes being required over the .osm file, and hence the splitter will take longer to run.
- There is no longer a limit on how many areas a way or relation can belong to (previously it was 4).