Manual Reference Pages  - formats (n)

NAME

formats - file formats of the Tina Toolbox

Part of Tina Toolbox for analysis of Petri nets and Time Petri nets.

CONTENTS

Description
The .net format
The .ndr format
The .tpn format
The .pnml format
The .tts format
The .ktz format
The .aut format
The .bcg format
The .mec format
The .adr format
The .ltl format
The .mmc format
The .scn format
Examples
See Also
Authors

DESCRIPTION

This page describe the formats of files used by the Tina tools.

Time Petri nets formats, input of tina/sift/tedd/struct/plan/play/ndrio/nd:

net.txt        .net    Time Petri nets textual format ndr.txt        .ndr    Time Petri nets graphic format tpn.txt        .tpn    Time Petri nets script format pnml.txt        .pnml   Time Petri nets pnml format

Time Transition System format, input of tina/sift/tedd/struct/plan/play (not supported by ndrio nor nd):

tts.txt        .tts    Time Transition System format

Transition system formats, output of tina:

aut.txt        .aut    Transition systems textual format (CADP format) ktz.txt        .ktz    Kripke transition systems binary format (tina) bcg.txt        .bcg    Transition systems binary format (tina for CADP) mec.txt        .mec    Kripke transition systems textual format (tina for MEC4)

Other formats:

adr.txt        .adr    Transition systems graphic format (nd) ltl.txt        .ltl    SE-LTL model checker commands (selt) mmc.txt        .ltl    Mu-calculus model checker commands (muse) scn.txt        .scn    Format of firing sequences and firing schedules,                         (selt, plan, play, nd stepper history files)

The .net format

This is the textual description format of Time Petri nets.

A net is described by a series of declarations of places, transitions and/or notes, and an optional naming declaration for the net. The net described is the superposition of these declarations. The grammar of .net declarations is the following, in which nonterminals are bracketed by < .. >, terminals are in upper case or quoted. Spaces, carriage return and tabs act as separators.

Optionally, labels may be assigned to places and transitions. This should be preferably done within "tr" and "pl" declarations rather than using separate "lb" declarations. The later form ("lb") is kept for backward compatibility and might disappear in future releases.

Grammar:

.net                    ::= (<trdesc>|<pldesc>|<lbdesc>|<prdesc>|<ntdesc>|<netdesc>)*
netdesc                 ::= ’net’ <net>
trdesc                  ::= ’tr’ <transition> {":" <label>} {<interval>} {<tinput> -> <toutput>}
pldesc                  ::= ’pl’ <place> {":" <label>} {(<marking>)} {<pinput> -> <poutput>}
ntdesc                  ::= ’nt’ <note> (’0’|’1’) <annotation>
lbdesc                  ::= ’lb’ [<place>|<transition>] <label>
prdesc                  ::= ’pr’ (<transition>)+ ("<"|">") (<transition>)+
interval                        ::= (’[’|’]’)INT’,’INT(’[’|’]’) | (’[’|’]’)INT’,’w[’
tinput                  ::= <place>{<arc>}
toutput                 ::= <place>{<normal_arc>}
pinput                  ::= <transition>{<normal_arc>}
poutput                 ::= <transition>{arc}
arc                     ::= <normal_arc> | <test_arc> | <inhibitor_arc> |
                            <stopwatch_arc> | <stopwatch-inhibitor_arc>
normal_arc              ::= ’*’<weight>
test_arc                ::= ’?’<weight>
inhibitor_arc           ::= ’?-’<weight>
stopwatch_arc           ::= ’!’<weight>
stopwatch-inhibitor_arc ::= ’!-’<weight>
weight, marking         ::= INT{’K’|’M’|’G’|’T’|’P’|’E’}
net, place, transition, label, note, annotation ::= ANAME | ’{’QNAME’}’
INT                     ::= unsigned integer
ANAME                   ::= alphanumeric name, see Notes below
QNAME                   ::= arbitrary name, see Notes below

Notes:

Two forms are admitted for net, place and transition names:
- ANAME : any non empty string of letters, digits, primes ’ and underscores _
- ’{’QNAME’}’ : any chain between braces, and in which characters {, }, and \ are prefixed by \

Empty lines and lines beginning with ’#’ are considered comments.

In any closed temporal interval [eft,lft], one must have eft <= lft.

The letter ’K’ (resp. ’M’, ’G’, ’T’, ’P’, ’E’) following a weight or marking multiplies it by 10^3 (resp. 10^6, 10^9, 10^12, 10^15, 10^18).

Weight is optional for normal arcs, but mandatory for test and inhibitor arcs

By default:
- transitions have temporal interval [0,w[
- normal arcs have weight 1
- places have marking 0
- places and transitions have the empty label "{}"

When several labels are assigned to some node, only the last assigned is kept.

The .ndr format

Format ndr is the format of graphic files produced by the nd editor. A net is described by a series of declarations of places, transitions and edges, followed by a net name declaration. The net described is the superposition of these declarations.

Grammar :

.ndr            ::= (<trdesc>|<pldesc>)* (<edgedesc>|<prdesc>|<ntdesc>)* <netdesc>
trdesc          ::= ’t’ <xpos> <ypos> <transition> <eft> <lft> <anchor>
                 |  ’t’ <xpos> <ypos> <transition> <anchor> <eft> <lft> <anchor> <label> <anchor>
pldesc          ::= ’p’ <xpos> <ypos> <place> <marking> <anchor> {<label> <anchor>}
ntdesc          ::= ’n’ <xpos> <ypos> <note> (’0’|’1’) <annotation>
edgedesc        ::= ’e’ <place> <transition> {<arckind>}<weight> <anchor>
                 |  ’e’ <place> <ang> <rad> <transition> <ang> <rad> {<arckind>}<weight> <anchor>
                 |  ’e’ <transition> <place> <weight> <anchor>
                 |  ’e’ <transition> <ang> <rad> <place> <ang> <rad> <weight> <anchor>
prdesc          ::= ’e’ <transition> <transition> 1 <anchor>
                 |  ’e’ <transition> <ang> <rad> <transition> <ang> <rad> 1 <anchor>
netdesc         ::= ’h’ <net> {<nodesize> {<bgcolor>}}
eft             ::= {-}INT
lft             ::= {-}INT | ’w’
weight, marking ::= INT{’K’|’M’|’G’|’T’|’P’|’E’}
arckind         ::= ’?’                              // test (read) arc
                 |  ’?-’                             // inhibitor
                 |  ’!’                              // stopwatch
                 |  ’!-’                             // stopwatch-inhibitor
xpos, ypos, rad ::= FLOAT
ang             ::= UFLOAT
net, place, transition, label, note, annotation ::= ANAME | ’{’QNAME’}’
anchor          ::= ’n’ | ’nw’ | ’w’ | ’sw’ | ’s’ | ’se’ | ’e’ | ’ne’ | ’c’
FLOAT           ::= unsigned float (without exponent)
UFLOAT          ::= unsigned float between 0 and 1 (without exponent)
INT             ::= unsigned integer
ANAME           ::= see notes below
QNAME           ::= see notes below
<nodesize>      ::= ’small’ | ’normal’ | ’large’
<bgcolor>       ::= any tcl-tk color

Notes:

Node declarations must precede edge declarations.

The last declaration must be the netname declaration (h).

Empty lines and lines beginning with ’#’are considered comments.

A ’-’ starting an eft or lft denotes an open interval end.

In any transition declaration, one must have ‘lft >= ‘eft or lft = ’w’, where ‘e denotes e if e>=0, or (-e)-1 otherwise

If eft = lft, then they may not be both negative (intervals may not be empty).

The letter ’K’ (resp. ’M’, ’G’, ’T’, ’P’, ’E’) following a weight or marking multiplies it by 10^3 (resp. 10^6, 10^9, 10^12, 10^15, 10^18).

Two forms are admitted for net, place and transition names:
- ANAME : any non empty string of letters, digits, primes ’ and underscores _
- ’{’QNAME’}’ : any chain between braces, and in which characters {, }, and \ are prefixed by \

The .tpn format

Format .tpn is a preliminary script language for building nets from net components specified in .net or .ndr format.

Syntax of tpn files:

A .tpn file is constituted of zero or more lines

        .tpn ::= <tpnline>*

each line is either any line allowed in a .net or .ndr file, or a tpn command:

        tpnline ::=
                  <trdesc>      (as in .net or .ndr formats)
                | <pldesc>      (as in .net or .ndr formats)
                | <lbdesc>      (as in .net or .ndr formats)
                | <prdesc>      (as in .net or .ndr formats)
                | <ntdesc>      (as in .net or .ndr formats)
                | <netdesc>     (as in .net or .ndr formats)
                | <tpncmd>
        tpncmd ::=
                  ’new’
                | ’dup’ INT
                | ’load’ TPNFILE
                | ’ren’ <renlist>
                | ’merge’ INT
                | ’sync’ INT
                | ’chain’ INT
                | ’source’ FILE
        renlist ::= <labpair> <renlist> | <labpair>
        labpair ::= <label>"/"<label> | "/"<label>
        label ::= a label (as in .net or .ndr formats)
        TPNFILE ::= the name of a .net or .ndr file
        FILE ::= the name of a .net, .ndr, or .tpn file

Note that, syntactically, .net and .ndr files are .tpn files. The tpn format extends both the .net and .ndr formats, but it has no graphic interpretation at the moment (a future .tdr format might provide that).

Interpretation of tpn scripts:

A tpn file describes a Time Petri net, possibly resulting from composition of several Time Petri nets.

Tpn scripts are interpreted as code for an abstract stack machine. The stack initially holds the empty Time Petri net. The result of evaluation of a tpn script is the net on top of the stack at completion of interpretation.

The lines of a tpn file are interpreted as follows ("top" means the net in topmost position on stack):

new                     pushes an empty net on stack
.net or .ndr line       extends top with the .net or .ndr declaration
dup n                   pushes on stack n copies of top (default n is 1)
ren R1 ... Rn           applies relabelling R1 ... Rn to top
merge n                 replaces n topmost nets by their concurrent composition
sync n                  replaces n topmost nets by their synchronization
chain n                 replaces n topmost nets by their chaining
source file             evaluates commands from file tpnfile
load tpnfile            pushes the empty net on stack, then sources tpnfile

Notes:

- merge n, sync n, chain n, assign unique names to the nodes of their components;

- synchronization (sync) fuse copies of transitions with same label in all components;

- chaining (chain) is like synchronization, but on places;

- concurrent composition (merge) is simply juxtaposition, preserving labels.
It is a derived form that could always be replaced by some combination of
relabelling and sync/chain (see examples);

- in a ren specification, all renamings and/or hidings are applied simultaneously;
ren x/y means that all nodes labelled by y become labelled by x
ren /x means that nodes labelled x become unlabelled (hidden)

- "load tpnfile" is a shorthand for the two lines:

        new
        source tpnfile

- the file names in source and load instructions are relative to the directory
in which resides the script file;

The .pnml format

The PNML format is still evolving. The tina tools currently accept BasicPNML (http://www.informatik.hu-berlin.de/top/pnml/basicPNML.rng) and an extension of it, TpnPNML, informally explained below.

TpnPNML extends BasicPNML as follows:

- In the "graphics" element of any node, the following line can be added,
that specifies an offset for the node identifier:

<offset x=... y=... />

- Any place and transition can have a "label" element, in addition to a
name element.

- Any transition can have a "delay" element, encoding a firing interval,
and possibly an offset for this interval. The syntax of intervals is
taken from MathML (http://www.w3.org/1998/Math/MathM).

  For instance, an interval [4,9] with offset (-10,0) is specified by:

<delay> <interval xmlns="http://www.w3.org/1998/Math/MathML" closure="closed"> <cn>4</cn> <cn>9</cn> </interval> <graphics> <offset x="-10" y="0" /> </graphics> </delay>

The unbounded interval [4,w[ would be specified by:

<delay> <interval xmlns="http://www.w3.org/1998/Math/MathML" closure="closed-open"> <cn>4</cn> <ci>infty</ci> </interval> </delay>

And interval ]3,5[ by:

<delay> <interval xmlns="http://www.w3.org/1998/Math/MathML" closure="open"> <cn>3</cn> <cn>5</cn> </interval> </delay>

Tools tina, struct and plan accept BasicPNML and TpnPNML natively.

The nd tool can import or export nets into TpnPNML. If the file has extension .pnml, then "file->open" and "file->save as" can be used for this, otherwise use "file->import" or "file->export".

The ndrio tool is able to convert .net and .ndr to/from .pnml.

The .tts format

As of 2.9.0, tina accepts as input high level descriptions called "Time Transition Systems" (TTS). A Time Transition System description consists of a Time Petri net described in any format understood by tina (.net, .ndr, .tpn, .pnml) and a shared library (a .so/.dyn/.dll file) defining the data processing synchronized with the net transitions, together in a directory with extension .tts.

Applied to a .tts directory, tina loads the .net or .ndr file (as usual), and dynamically links the library file to get the actions and predicates. This file described the API the shared library must obey.

tina sees data values as keys (integers, not necessarily contiguous), through a "store" function. The storage functions may reside in a separate library or may be statically linked with the data library, all storage libraries must provide functions:

        store : value -> key
        lookup : key -> value
        init_storage : unit -> unit

The data shared library itself captures:

    an array of transition names for the transitions referred to
    in the library. The size of the array is returned in the parameter
    of function transition.

        transitions : int ref -> string vector

A function returning the current version of the format (currently 2):

ttsapiversion : unit -> int

A function initial returning the initial data state and a function final called after exploration is complete:

initial : unit -> key final : int -> unit

for a subset of transitions in transitions, referred to by their index in array transitions, functions:

        pre_i : key -> bool         act_i : key -> key

A function set_exceptions, called by the tts client if wishing to handle tts runtime exceptions itself rather then leaving that task to the tts library (default behavior). Function set_exception returns the array of possible exception messages, its size in the first argument, and, in the second argument, the address of the C variable holding the exception code when some exception is raised. Exception codes are either an exception number in 0..99 or a line number (line) and an exception number (no) encoded into the integer 100*line + no. To handle exceptions, the client must read that variable after each call of pre_i and act_i, and reset it to 0 afterwards:     set_exceptions : int ref * pointer ref -> string vector

An independance predicate for actions. This predicate is used by tina for computing the conflicting relation for partial order methods, and the persistent and newly enabled relations for state class graphs:

independant : int * int -> bool

A value printer, used by tina to print data value in some concrete notation when output is "verbose":

        sprint_state: key * string ref -> int

For kts output, the state (or some abstraction of it) should be available to tina as a record with integer fields. This is done by providing:

A vector of observation labels, and the number of, by function:

        obs_names : int ref -> string vector

A function to retrieve the contents of observables from a value:

        obs_values : key -> int vector

A first high level description language compiled into this format, called Fiacre, should be available by now.

The .ktz format

The ktz format is the binary format for kripke transition system (kts) of tina. That compact format is the input format of selt (the tina SE-LTL model checker) and other forthcoming tools.

(description to be provided).

The .aut format

The aut format is the lts description format of the Aldebaran tool of the CADP toolset (see the CADP documents).

Notes

The .aut format does not allow double quotes (") to appear in labels. In addition tina forbids backquotes (‘) in .aut labels. Backquotes are used by tina to encode sets of properties into single .aut labels.

The .bcg format

The bcg format is the binary lts description format of the BCG tools of the CADP toolset (see the CADP documents).

Notes

tina forbids double quotes (") and backquotes (‘) to appear in .bcg labels. Backquotes are used by tina to encode lists of properties into single .bcg labels; double quotes would prevent lossless conversions to .bcg.

The .mec format

The mec format is the kts description format of the Mec4 model checker.

Notes

mec only allows alphanumeric labels (a letter followed by letters, digits or _) to appear as state or transition properties. tina will refuse to generate .mec descriptions if place or transition names are not legal mec property identifiers.

The .adr format

It is the format of graphic files produced by the nd editor when editing automata. An automaton is described by a series of declarations of states and edges, followed by a name declaration. The automaton described is the superposition of these declarations.

Grammar:

.adr                    ::= (<stdesc>)* (<edgedesc>)* <autdesc>
stdesc                  ::= ’s’ <xpos> <ypos> <state> {<anchor> <label> <anchor>}
edgedesc                ::= ’e’ <state> <state> <label> <anchor>
edgedesc                 |  ’e’ <state> <ang> <rad> <state> <ang> <rad> <label> <anchor>
autdesc                 ::= ’h’ <aut> <attr>
xpos, ypos, rad         ::= FLOAT
ang                     ::= UFLOAT
aut, state, label       ::= ANAME | ’{’QNAME’}’
anchor                  ::= ’n’ | ’nw’ | ’w’ | ’sw’ | ’s’ | ’se’ | ’e’ | ’ne’ | ’c’
FLOAT                   ::= integer or float (without exponent)
UFLOAT                  ::= integer or float between 0 and 1 (without exponent)
INT                     ::= unsigned integer
ANAME                   ::= see notes below
QNAME                   ::= see notes below
attr                    ::= an unspecified number of optional graphic parameters (specifying size, color, etc)

Notes:

State declarations must precede edge declarations.

The last declaration must be the autdesc declaration (h).

Empty lines and lines beginning with ’#’are considered comments.

Two forms are admitted for net, place and transition names:
- ANAME : any non empty string of letters, digits, primes ’ and underscores _
- ’{’QNAME’}’ : any chain between braces, and in which characters {, }, and \ are prefixed by \

The .ltl format

ltl is the format of files accepted as use input by the selt model-checker. Check the selt manual page for a full description.

The .mmc format

mmc is the format of files accepted as use input by the muse model-checker. Check the muse manual page for a full description.

The .scn format

The scn format is used by:

- the nd stepper to store/load histories;
- the selt model checker to save counter examples for replay by the stepper;
- the plan application, to analyse path timing.

A .scn file is a series of line, each constituted of a series of blocks of one of the following shape (all blocks in a file have same shape) :

       <transition>
  or:  <transition>@<time>
  or:  <transition>$<time>

where <transition> is a transition name (see net.txt) and <time> is a nonnegative float (without exponent)

Spaces and tabs are considered separators;

Line breaks are significant, they are interpreted as marks by the nd stepper, and preserved by the plan tool;

Blank lines and and those beginning with "#" are considered comments;

EXAMPLES

Directory "net" in tina distributions include various example descriptions in the above formats.

SEE ALSO

nd(n), tina(n), plan(n), struct(n), ktzio(n), ndrio(n), tedd(n), selt(n), sift(n), muse(n), pathto(n), scan(n), play(n), walk(n), reduce(n)

AUTHORS

Bernard Berthomieu, LAAS/CNRS, 2000-2024, Bernard.Berthomieu@laas.fr.


Tina Formats formats (n) Version 3.8.0
Generated by manServer 1.07 from src/formats.n using man macros.