#Please insert up references in the next lines (line starts with keyword UP)
UP      arb.hlp
UP      glossary.hlp

#Please insert subtopic references  (line starts with keyword SUB)
SUB     parser.hlp
SUB     commands.hlp

# Hypertext links in helptext can be added like this: LINK{ref.hlp|http://add|bla@domain}

#************* Title of helpfile !! and start of real helpfile ********
TITLE           How to define new import formats

OCCURRENCE      ARB_NT

SECTION         BRIEF DESCRIPTION

                All description files of import formats are located in
                the directory '$ARBHOME/lib/import'.
                Each of these files describe how to analyze the input files.

                A basic import description file (.ift) looks like this:

                        [AUTODETECT     "Matchpattern"]
                        BEGIN           "Matchpattern"
                        [KEYWIDTH       #Columnnumber]
                        [AUTOTAG        ["TAGNAME"]]
                        [IFNOTSET       x "Reason why x is not set"]+
                        [SETGLOBAL      x "global value"]+
                        [INCLUDE        "file"]+
                        [MATCH          "Matchpattern"
                                [SRT            "SRT_STRING"]
                                [ACI            "ACI_STRING"]
                                [TAG            "TAGNAME"]
                                [WRITE          "DB_FIELD_NAME"]
                                [WRITE_INT      "DB_FIELD_NAME"]
                                [WRITE_FLOAT    "DB_FIELD_NAME"]
                                [APPEND         "DB_FIELD_NAME"]
                                [SETVAR         x]
                        ]*
                        SEQUENCESTART   "Matchpattern"
                        SEQUENCECOLUMN  #Columnnumber
                        [SEQUENCESRT    "SRT_STRING"]
                        [SEQUENCEACI    "ACI_STRING"]
                        SEQUENCEEND     "STRING"
                        [CREATE_ACC_FROM_SEQUENCE]
                        [DONT_GEN_NAMES]
                        END             "STRING"

                or it can pipe the data through any external program PROGRAM to
                convert it to an already existing format 'exformat'
                using the following basic design:

                [AUTODETECT     "Matchpattern"]
                SYSTEM          "PROGRAM $< $>"
                NEW_FORMAT      "lib/import/exformat.ift"

                $< will be replaced by the input file name
                $> will ve replaced by the intermediate file name

DESCRIPTION     First of all the converter appends all import files maching
                the filepattern into one file. The files are separated by the
                string defined with the keyword  SEQUENCEEND.

                1. Search the first line matching the pattern defined by BEGIN

                2. Try to match all MATCH_patterns.

                   For all lines that match do:

                        - append all following lines, which start after
                          column KEYWIDTH

                        - run commands with the concatenated lines

                   Known commands are (they are executed in the order listed here):

                         - SRT "SRT_STRING"

                               start the string replace tool on the current result and
                               set the output as current result.

                         - ACI "ACI_STRING"

                               run the arb command interpreter to change the current result.

                         - TAG "TAGNAME"

                               tag information (i.e. "[EBI] 1997 [RDP] 1998")

                         - WRITE "DB_FIELD_NAME"

                               write the current result into DB_FIELD_NAME 

                         - WRITE_INT "DB_FIELD_NAME"

                               like WRITE, but expect integer target field 

                         - WRITE_FLOAT "DB_FIELD_NAME"

                               like WRITE, but expect floating-point target field 

                         - APPEND "DB_FIELD_NAME"

                               append the current result to DB_FIELD_NAME

                         - SETVAR x

                               store the current result in the variable x, where x may be any character.
                               After it was set this variable can be referenced by using $x in any
                               command expression (SRT_STRING,ACI_STRING,TAGNAME,DB_FIELD_NAME).

                               For each used variable there has to be defined an error reason
                               describing, what's wrong if the variable has NOT been set.
                               Define error reasons using

                                      IFNOTSET x "Reason why x is not set"

                               Note: use '$$' to insert a single '$'.

                               Allowed variable names are 'a' to 'z'.

                   Note: Every of these commands may only occur once in one MATCH rule.
                         To run some of them multiple, create multiple MATCH rules.

                3. If the line matches SEQUENCESTART_pattern, assume that
                   all following lines to and except the line
                   matching SEQUENCEEND_pattern contain the sequence data.

                4. GOTO 1

                Postprocesses:

                        CREATE_ACC_FROM_SEQUENCE:

                                Generate a checksum for all sequences with no accession
                                entry ('acc' -field) and write it as the accession number

                        DONT_GEN_NAMES:

                                Do not try to generate unique short names for the species
                                using the full_name field.

                General commands:

                        INCLUDE "filename"

                                Simply inserts the contents of "filename" at the current position.

                                It's possible to declare variables in the file where the INCLUDE
                                happens and to use them in the included file. (Example: 
                                longebi.ift, longgenbank.ift and feature_table.ift in subdir nonformats)

                        SETGLOBAL x "value"

                                Sets global variable 'x' to 'value'.

                        AUTOTAG ["TAGNAME"]

                                If set, act like each MATCH rule has a
                                   TAG "TAGNAME"
                                entry. Use AUTOTAG w/o parameter to reset
                                to default behavior.

EXAMPLES        Look at the files in '$ARBHOME/lib/import'

WARNINGS        Format detection does not always work