Using the GEDCOM parser library
Index
Overview
The GEDCOM
parser library provides two interfaces. At the one hand, it can be
used as a callback-based parser (comparable to the SAX interface of
XML); at the other hand, the parser can be used to convert the GEDCOM file
into an object model (comparable to the DOM interface of XML). It comes
with:
- a library (
libgedcom.so
), to be linked in the
application program, which implements the callback parser
- a header file (
gedcom.h
), to be used in the sources
of the application program
- a header file (
gedcom-tags.h
) that is also installed,
but that is automatically included via gedcom.h
Additionally, if you want to use the GEDCOM C object model, the following should be used (note that libgedcom.so
is also needed in this case, because the object model uses the callback parser internally):
- a library (
libgedcom_gom.so
), to be linked in the application program, which implements the C object model
- a header file (
gom.h
), to be used in the sources of the application program
There is a separate script to help with library and compilation flags, see the development support.
Next to these, there is also a data directory in $PREFIX/share/gedcom-parse
that contains some additional stuff, but which is not immediately
important at first. I'll leave the description of the data directory
for later.
The very simplest call of the gedcom callback parser is simply the following
piece of code (include of the gedcom.h
header is assumed, as everywhere
in this manual):
int result;
...
gedcom_init();
...
result = gedcom_parse_file("myfamily.ged");
Although this will not provide much information, one thing it does
is parse the entire file and return the result. The function returns
0 on success and 1 on failure. No other information is available
using this function only.
Alternatively, programs using the C object model should use the following (in this case, the inclusion of both gedcom.h
and gom.h
is required):
int result;
...
gedcom_init();
...
result = gom_parse_file("myfamily.ged");
The call to gom_parse_file
will build the C object model, which is then a complete representation of the GEDCOM file.
No matter which of the interfaces you use, the call to gedcom_init
() should be one of the first calls
in your program. The requirement is that it should come before the first
call to iconv_open
(part of the generic character set conversion
feature) in the program, either by your program itself, or indirectly by
the library calls it makes. Practically, it should e.g. come before
any calls to any GTK functions, because GTK uses iconv_open
in its initialization.
For the same reason it is also advised to put
the -lgedcom
option
on the linking of the program as the last option, so that its initialization
code is run first. In the case of using the C object model, the linking
options should be: -lgedcom_gom -lgedcom
The function gedcom_init()
also initializes locale handling by calling setlocale(LC_ALL, "")
, in case the application would not do this (it doesn't hurt for the application to do the same).
The next sections will refine this piece of code to be able to have
meaningful errors and the actual data that is in the file.
Error handling
The library can be used in several different circumstances, both
terminal-based as GUI-based. Therefore, it leaves the actual display
of the error message up to the application. For this, the application
needs to register a callback before parsing the GEDCOM file, which will
be called by the library on errors, warnings and messages.
A typical piece of code would be (gom_parse_file
would be called in case the C object model is used):
void my_message_handler (Gedcom_msg_type type,
char *msg)
{
...
}
...
gedcom_set_message_handler(my_message_handler);
...
result = gedcom_parse_file("myfamily.ged");
In the above piece of code, my_message_handler
is the
callback that will be called for errors (type=ERROR
), warnings
(type=WARNING
) and messages (type=MESSAGE
). The
callback must have the signature as in the example. For errors,
the msg
passed to the callback will have the format:
Error on line
<lineno>: <actual_message>
Note that the entire string will be properly internationalized, and
encoded in UTF-8 (Why UTF-8?). Also,
no newline is appended, so that the application program can use it in any
way it wants. Warnings are similar, but use "Warning" instead of "Error".
Messages are plain text, without any prefix.
With this in place, the resulting code will already show errors and
warnings produced by the parser, e.g. on the terminal if a simple
printf
is used in the message handler.
Data callback mechanism
The most important use of the parser is of course to get the data
out of the GEDCOM file. This section focuses on the callback mechanism (see here for the C object model). In fact, the mechanism involves two levels.
The primary level is that each of the sections in a GEDCOM file is
notified to the application code via a "start element" callback and an
"end element" callback (much like in a SAX interface for XML), i.e. when
a line containing a certain tag is parsed, the "start element" callback
is called for that tag, and when all its subordinate lines with their
tags have been processed, the "end element" callback is called for the
original tag. Since GEDCOM is hierarchical, this results in properly
nested calls to appropriate "start element" and "end element" callbacks (note: see compatibility handling).
However, it would be typical for a genealogy program to support only
a subset of the GEDCOM standard, certainly a program that is still under
development. Moreover, under GEDCOM it is allowed for an application
to define its own tags, which will typically not be supported by
another application. Still, in that case, data preservation is important;
it would hardly be accepted that information that is not understood by
a certain program is just removed.
Therefore, the second level of callbacks involves a "default callback".
An application needs to subscribe to callbacks for tags it does support,
and need to provide a "default callback" which will be called for tags
it doesn't support. The application can then choose to just store
the information that comes via the default callback in plain textual format.
After this introduction, let's see what the API looks like...
Start and end callbacks
Callbacks for records
As a simple example, we will get some information from the header
of a GEDCOM file. First, have a look at the following piece of code:
Gedcom_ctxt my_header_start_cb (Gedcom_rec rec,
int level,
Gedcom_val xref,
char *tag,
char *raw_value,
int parsed_tag,
Gedcom_val parsed_value)
{
printf("The header starts\n");
return (Gedcom_ctxt)1;
}
void my_header_end_cb (Gedcom_rec rec, Gedcom_ctxt self)
{
printf("The header ends, context is %d\n", (int)self);
/* context will print as "1" */
}
...
gedcom_subscribe_to_record(REC_HEAD, my_header_start_cb,
my_header_end_cb);
...
result = gedcom_parse_file("myfamily.ged");
Using the gedcom_subscribe_to_record
function, the
application requests to use the specified callbacks as start and end
callback. The end callback is optional: you can pass NULL
if you are not interested in the end callback. The identifiers
to use as first argument to the function (here REC_HEAD
)
are described in the interface
details . These are also passed as first argument in the callbacks (the Gedcom_rec
argument).
From the name of the function it becomes clear that this function
is specific to complete records. For the separate elements in records
there is another function, which we'll see shortly. Again, the callbacks
need to have the signatures as shown in the example.
The Gedcom_ctxt
type that is used as a result of the
start callback and as an argument to the end callback is vital for passing
context necessary for the application. This type is meant to be opaque;
in fact, it's a void pointer, so you can pass anything via it. The
important thing to know is that the context that the application returns
in the start callback will be passed in the end callback as an argument,
and as we will see shortly, also to all the directly subordinate elements
of the record.
The tag
is the GEDCOM tag in string format, the parsed_tag
is an integer, for which symbolic values are defined as TAG_HEAD,
TAG_SOUR,
TAG_DATA,
... and USERTAG
for the application-specific tags. These values
are defined in the header gedcom-tags.h
that is installed,
and included via gedcom.h
(so no need to include gedcom-tags.h
yourself).
The example passes a simple integer as context, but an application
could e.g. pass a struct
(or an object in a C++ application)
that will contain the information for the header. In the end callback,
the application could then e.g. do some finalizing operations on the
struct
to put it in its database.
(Note that the Gedcom_val
type for the xref
and parsed_value
arguments was not discussed, see further
for this)
Callbacks for elements
We will now retrieve the SOUR field (the name of the program that
wrote the file) from the header:
Gedcom_ctxt my_header_source_start_cb(Gedcom_elt elt,
Gedcom_ctxt
parent,
int
level,
char*
tag,
char*
raw_value,
int
parsed_tag,
Gedcom_val
parsed_value)
{
char *source = GEDCOM_STRING(parsed_value);
printf("This file was written by %s\n", source);
return parent;
}
void my_header_source_end_cb(Gedcom_elt elt,
Gedcom_ctxt parent,
Gedcom_ctxt self,
Gedcom_val parsed_value)
{
printf("End of the source description\n");
}
...
gedcom_subscribe_to_element(ELT_HEAD_SOUR,
my_header_source_start_cb,
my_header_source_end_cb);
...
result = gedcom_parse_file("myfamily.ged");
The subscription mechanism for elements is similar, only the signatures
of the callbacks differ. The signature for the start callback shows
that the context of the parent line (here e.g. the struct
that describes the header) is passed to this start callback. The
callback itself returns here in this example the same context, but this
can be its own context object of course. The end callback is called
with both the context of the parent and the context of itself, which in this
example will be the same. Again, the list of identifiers to use as
a first argument for the subscription function are detailed in the interface details . Again, these are passed as first argument in the callback (the Gedcom_elt
argument).
If we look at the other arguments of the start callback, we see the
level number (the initial number of the line in the GEDCOM file), the tag
(e.g. "SOUR"), and then a raw value, a parsed tag and a parsed value. The
raw value is just the raw string that occurs as value on the line next
to the tag (in UTF-8 encoding). The parsed value is the meaningful
value that is parsed from that raw string. The parsed tag is described
in the section for record callbacks above.
The Gedcom_val
type is meant to be an opaque type. The
only thing that needs to be known about it is that it can contain specific
data types, which have to be retrieved from it using pre-defined macros.
These data types are described in the interface details.
Some extra notes:
- The
Gedcom_val
argument of the end callback
is used to pass 'complete' values, e.g. the full text of a note. See
the interface details
for the exact type of this argument.
- There are also two
Gedcom_val
arguments
in the start callback for records. The first one (xref
) contains the xref_value
corresponding to the cross-reference
(or NULL
if there isn't one), the second one (parsed_value
) contains the value that is parsed from the raw_value
. See
the interface details
.
Default callbacks
As described above, an application doesn't always implement the entire
GEDCOM spec, and application-specific tags may have been added by other
applications. To preserve this extra data anyway, a default callback
can be registered by the application, as in the following example:
void my_default_cb (Gedcom_elt elt, Gedcom_ctxt parent, int level,
char* tag, char* raw_value, int parsed_tag)
{
...
}
...
gedcom_set_default_callback(my_default_cb);
...
result = gedcom_parse_file("myfamily.ged");
This callback has a similar signature as the previous ones,
but it doesn't contain a parsed value. However, it does contain the
parent context, that was returned by the application for the most specific
containing tag that the application supported.
Suppose e.g. that this callback is called for some tags in the header
that are specific to some other application, then our application could
make sure that the parent context contains the struct or object that represents
the header, and use the default callback here to add the level, tag and
raw_value as plain text in a member of that struct or object, thus preserving
the information. The application can then write this out when the
data is saved again in a GEDCOM file. To make it more specific, consider
the following example:
struct header {
char* source;
...
char* extra_text;
};
Gedcom_ctxt my_header_start_cb(Gedcom_rec rec, int level, Gedcom_val xref, char* tag,
char *raw_value,
int parsed_tag, Gedcom_val parsed_value)
{
struct header head = my_make_header_struct();
return (Gedcom_ctxt)head;
}
void my_default_cb(Gedcom_elt elt, Gedcom_ctxt parent, int level, char* tag, char*
raw_value, int parsed_tag)
{
struct header head = (struct header)parent;
my_header_add_to_extra_text(head, level, tag, raw_value);
}
gedcom_set_default_callback(my_default_cb);
gedcom_subscribe_to_record(REC_HEAD, my_header_start, NULL);
...
result = gedcom_parse_file(filename);
Note that the default callback will be called for any tag that isn't
specifically subscribed upon by the application, and can thus be called
in various contexts. For simplicity, the example above doesn't take
this into account (the parent
could be of different
types, depending on the context).
Note also that the default callback is not called when the parent context
is NULL
. This is e.g. the case if none
of the "upper" tags has been subscribed upon.
Support for writing GEDCOM files
The Gedcom parser library also contains functions to writing GEDCOM files.
Similar as for the parsing itself, there are two interfaces: an interface
which is very basic, and requires you to call a function for each line in
the GEDCOM file, and an interface which just dumps the Gedcom object model
to a file in one shot (if you use the Gedcom object model).
Again, this section focuses on the basic interface, the Gedcom object model interface is described here.
Opening and closing files
The basic functions for opening and closing Gedcom files for writing are the following:
Gedcom_write_hndl gedcom_write_open (const char* filename);
int gedcom_write_close (Gedcom_write_hndl hndl, int* total_conv_fails);
The function gedcom_write_open
takes a parameter the name of
the file to write, and returns a write handle, which needs to be used in
subsequent functions. It returns NULL
in case of errors.
The function gedcom_write_close
takes, next to the write handle,
an integer pointer as parameter. If you pass an actual pointer for
this, the function will write in it the total number of conversion failures;
you can pass NULL
if you're not interested. The function returns 0 in case of success, non-zero in case of failure.
Controlling some settings
Note that by default the file is written in the same encoding as the read file was in. You can change
this by calling the following function before calling gedcom_write_open
, i.e. it affects all files that are opened after it is being called:
int gedcom_write_set_encoding (Enc_from from, const char* charset, Encoding width, Enc_bom bom);
The from
parameter indicates how you want the encoding to be set:
ENC_FROM_FILE
: The same as the read file was in (this is the default).
ENC_FROM_SYS
: Not a valid value here, see below for gedcom_write_set_terminator
ENC_MANUAL
: From the values given in the following parameters.
When ENC_FROM_FILE
is selected, the other parameters in the function are ignored (they can be passed as 0). When ENC_MANUAL
is chosen, the meaning of the other parameters is as follows:
The valid charset
values are given in the first column in the file gedcom.enc
in the data directory of gedcom-parse ($PREFIX/share/gedcom-parse
).
The character sets UNICODE, ASCII and ANSEL are always supported (these
are standard for GEDCOM), as well as ANSI (not standard), but there may be
others.
The width
parameter takes one of the following values:
ONE_BYTE
: This should be used for all character sets except UNICODE.
TWO_BYTE_HILO
: High-low encoding for UNICODE (i.e. big-endian)
TWO_BYTE_LOHI
: Low-high encoding for UNICODE (i.e. little-endian)
The bom
parameter determines whether a byte-order-mark should
be written in the file in case of UNICODE encoding (usually preferred because
it then clearly indicates the byte ordering). It takes one of the following
values:
For both these parameters you can pass 0 for non-UNICODE encodings,
since that corresponds to the correct values (and is ignored anyway). The
function returns 0 in case of success, non-zero in case of error. Note
that you still need to pass the correct charset value for the HEAD.CHAR tag,
otherwise you will get a warning, and the value will be forced to the correct
value.
Further, it is possible to control the kind of line terminator that is used, via the following function (also to be used before gedcom_write_open
):
int gedcom_write_set_line_terminator (Enc_from from, Enc_line_end end);
The values for the from
parameter are given above. The value ENC_FROM_SYS
is valid here, and means that the normal terminator for the current system
is used (the second parameter of the function is then ignored). This
is the default for this setting.
The end
parameter takes one of the following values:
END_CR
: only carriage return ("/r") (system value for Macintosh)
END_LF
: only line feed ("/n") (system value for Unix, Mac OS X)
END_CR_LF
: first carriage return, then line feed ("/r/n") (system value for DOS, Windows)
END_LF_CR
: first line feed, then carriage return ("/n/r")
By default, this is set to the appropriate line terminator on the current
platform, so it only needs to be changed if there is some special reason
for it.
Writing data
For actually writing the data, the principle is that every line in the GEDCOM
file to write corresponds to a call to one of the following functions, except
that CONT/CONC lines can be automatically taken care of. Note that
the resulting GEDCOM file should conform to the GEDCOM standard. Several
checks are built in already, and more will follow, to force this. There
is no compatibility mode for writing GEDCOM files (and probably never will be).
In general, each of the following functions expect their input in UTF-8 encoding (see also here). If this is not the case, errors will be returned.
Note that for examples of using these functions you can look at the sources for the Gedcom object model (e.g. the function write_header
in gom/header.c
).
Records
For writing lines corresponding to records (i.e. on level 0), the following function is available:
int gedcom_write_record_str (Gedcom_write_hndl hndl, Gedcom_rec rec, const char* xrefstr, const char* value);
The hndl
parameter is the write handle that was returned by gedcom_write_open
. The rec
parameter is one of the identifiers given in the first column in this table (except REC_USER
: see below). The xrefstr
and val
parameters are respectively the cross-reference key of the record (something like '@FAM01@
'), and the value of the record line, which should be NULL
for some record types, according to the same table.
Elements
For writing lines corresponding to elements (inside records, i.e. on a level
bigger than 0), the following functions are available, depending on the data
type:
int gedcom_write_element_str (Gedcom_write_hndl hndl, Gedcom_elt elt, int parsed_tag,
int parent_rec_or_elt, const char* value);
i
nt gedcom_write_element_xref (Gedcom_write_hndl hndl, Gedcom_elt elt, int parsed_tag,
int parent_rec_or_elt, const struct xref_value*
value);
int gedcom_write_element_date (Gedcom_write_hndl hndl, Gedcom_elt elt, int parsed_tag,
int parent_rec_or_elt, const struct date_value*
value);
i
nt gedcom_write_element_age (Gedcom_write_hndl hndl, Gedcom_elt elt, int parsed_tag,
int parent_rec_or_elt, const struct age_value*
value);
These functions only differ in the type of the last argument, which is the value of the element.
The hndl
parameter is again the write handle returned by gedcom_write_open
. The elt
parameter is one of the identifiers given in the first column in this table (except ELT_USER
: see below). The parent_rec_or_elt
is the corresponding rec
or elt
identifier of the logically enclosing statement: this will determine the
level number written on the line, as the level number of the parent + 1.
Some of the identifiers can actually stand for different tags. For this reason, the parsed_tag
has to be passed for some of them. This parsed tag is the same as was returned by the callback functions defined above, and is an identifier of the form TAG_name
. This parameter is needed whenever the second column in this table shows several possible tags (this is e.g. the case for ELT_SUB_FAM_EVT
).
Note that for writing a date value, the given value should be valid, i.e.
all its struct fields filled in properly and consistent. This can be
done by calling gedcom_normalize_date
(see here).
User-defined tags
For user-defined tags (tags starting with an underscore), there are separate functions, again depending on the data type:
int gedcom_write_user_str (Gedcom_write_hndl hndl, int level, const char* tag, const char* xrefstr,
const char* value);
i
nt gedcom_write_user_xref (Gedcom_write_hndl hndl,
int level, const char* tag, const char* xrefstr,
const struct xref_value* value);
In the case of user-defined tags, the level and tag string are passed verbatim
(not controlled by the library). This allows to write any extra data
that doesn't use a standard tag, but is only allowed for tags starting with
an underscore.
Other API functions
Although the above describes the basic interface of the gedcom parser, there
are some other functions that allow to customize the behaviour of the library.
These will be explained in the current section.
Debugging
The library can generate various debugging output, not only from itself,
but also the debugging output generated by the yacc parser. By default,
no debugging output is generated, but this can be customized using the
following function:
void gedcom_set_debug_level (int level, FILE*
trace_output)
The level
can be one of the following values:
- 0: no debugging information (this is the
default)
- 1: only debugging information from libgedcom
itself
- 2: debugging information from libgedcom
and yacc
If the trace_output
is NULL
, debugging information
will be written to stderr
, otherwise the given file handle
is used (which must be open).
Error treatment
One of the previous sections already described the callback to be
registered to get error messages. The library also allows to customize
what happens on an error, using the following function:
void gedcom_set_error_handling (Gedcom_err_mech
mechanism)
The mechanism
can be one of:
IMMED_FAIL
: immediately fail
the parsing on an error (this is the default)
DEFER_FAIL
: continue parsing
after an error, but return a failure code eventually
IGNORE_ERRORS
: continue parsing
after an error, return success always
This doesn't influence the generation of error or warning messages,
only the behaviour of the parser and its return code.
Compatibility mode
Applications are not necessarily true to the GEDCOM spec (or use a
different version than 5.5). The intention is that the library is
resilient to this, and goes in compatibility mode for files written by specific
programs (detected via the HEAD.SOUR tag). This compatibility mode
can be enabled and disabled via the following function:
void gedcom_set_compat_handling (int enable_compat)
The argument can be:
- 0: disable compatibility mode
- 1: allow compatibility mode (this is the
default)
Currently, there is (some) compatibility for:
- ftree
- Lifelines (3.0.2)
- Personal Ancestral File (PAF), version 2, 4 and 5
- Family Origins
- EasyTree
The following function allows to set some options for the compatibility handling:
void gedcom_set_compat_options (Gedcom_compat options)
The parameter can be an OR'ed combination of the following options:
COMPAT_ALLOW_OUT_OF_CONTEXT
In some compatibility cases, tags are coming out-of-order,
i.e. their start element callback would have to come after the end element
callback of the parent tag. E.g. instead of the standard GEDCOM
1 DATE ...
2 TIME ...
the genealogy program has generated something like:
1 DATE ...
1 TIME ...
This can give a problem if your end element callbacks free some resources.
If your program can handle elements out of context, you can enable this option.
By default it is disabled, and so the values of these out-of-context
tags are lost (the parser generates a warning if this is the case). Note:
currently the Gedcom object model in C has this option disabled too, although
this will change in the future.
Converting character sets
All strings passed by the GEDCOM parser to the application are in UTF-8
encoding. Typically, an application needs to convert this to something
else to be able to display it.
The most common case is that the output character set is controlled by
the locale
mechanism (i.e. via the LANG
,
LC_ALL
or LC_CTYPE
environment variables), which also
controls the gettext
mechanism in the application.
With
gedcom-parse
comes a library implementing help functions for UTF-8 encoding (
see
the documentation for this library).
Development support
Macro for configure.in
There
is a macro available for use in configure.in for applications that are using
autoconf to configure their sources. The following macro checks whether
the Gedcom parser library is available and whether its version is high enough:
AM_PATH_GEDCOM_PARSER([min_version,[action_if_found,[action_if_not_found,[modules]]]])
All the arguments are optional and default to 0. E.g. to check for
version 1.34.2, you would put in configure.in the following statement:
AM_PATH_GEDCOM_PARSER(1.34.2)
Note that version numbers now contains three parts (since version
0.20.0: this is also the first version in which this macro is available).
The macro also sets the variables GEDCOM_CFLAGS
and GEDCOM_LIBS
for use in Makefiles. Typically, this would be done as follows in a Makefile.am:
bin_programs = myprg
myprg_SOURCES = myprg.c foo.c bar.c
INCLUDES = @GEDCOM_CFLAGS@
LDADD = @GEDCOM_LIBS@
If your program uses some extra modules, they can be passed as fourth argument
in the macro, so that the CFLAGS and LIBS are correctly filled in. Currently,
the only available module is gom
(the Gedcom object model). For example:
AM_PATH_GEDCOM_PARSER(0.21.2, , ,gom)
To be able to use this macro in the sources of your application, you have three options:
- Put the file
m4/gedcom.m4
in your autoconf data directory (i.e. the path given by 'aclocal --print-ac-dir
', usually /usr/share/aclocal
). You can do this automatically by going into the m4 subdirectory and typing 'make install-m4
'.
- If you're using autoconf, but not automake, copy the contents of
m4/gedcom.m4
in the aclocal.m4
file in your sources.
- If you're using automake, copy the contents of
m4/gedcom.m4
in the acinclude.m4
file in your sources.
There are three preprocessor symbols defined for version checks in the
header (but their direct use is deprecated: please use the macro above):
GEDCOM_PARSE_VERSION_MAJOR
GEDCOM_PARSE_VERSION_MINOR
GEDCOM_PARSE_VERSION
The last one is equal to (GEDCOM_PARSE_VERSION_MAJOR * 1000) + GEDCOM_PARSE_VERSION_MINOR.
As you see, this only checked the major and minor version, not the patch number, so this is obsolete.
Compilation and linking flags
Similar to other libraries, the GEDCOM parse library installs a script gedcom-config
to help with compilation and linking flags for programs that don't use autoconf/automake.
To get compilation flags for your program, use (depending on whether you
only use the callback parser, or also the GEDCOM object model):
gedcom-config --cflags
gedcom-config --cflags gom
Similarly, to get linking flags, use one of the following:
gedcom-config --libs
gedcom-config --libs gom
$Id: usage.html,v 1.25 2003/02/02 14:40:06 verthezp Exp $
$Name: R0_90_0 $