/var/lib/mrs/parsers/00-readme.txt is in mrs 6.0.5+dfsg-3ubuntu1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 | Parser documentation for MRS version 6
1. Identify documents
Flat text databanks contain documents in some structured way. Sometimes it is
possible to locate a new document by looking at the first line, in other cases
documents are separated by a special string. And then databanks can have
leading or trailing pieces of text that should be ignored.
A parser script specifies the way to divide a text file into documents by
setting properties in the constructor for the parser. Recognized properties
are:
header, lastheaderline, trailer, firstdocline and lastdocline
For each of these properties you can assign a simple string in which case the
line should match exactly, or you can assign a regular expression and then
this expression is used to match a line.
If lastheaderline is defined, header is ignored.
If you did not provide any property, MRS assumes one document per file.
2. Parsing documents
The sub method of a parser is called with the document as second parameter.
It is the task of the parser to identify indexable text and pass it to MRS.
For this there are the functions:
index_text
Uses a tokenizer to split the text into words and numbers and stores
each along with their position in the document.
index_unique_string
Stores the passed string verbatim in an index. The value should be
unique, used for e.g. ID fields.
index_string
Stores the passed string verbatim, but it does not have to be unique.
index_number
Stores the number in an index that is sorted on numeric value rather
than the textual representation.
The parser can append meta data to a document using the set_attribute call.
Two attributes are strongly recommended, one is 'id' and the other is 'title'.
They are not strictly needed, but are very useful.
If links between the current document and another document can be found, the
link information can be stored using the add_link call. The first parameter
is the databank containing the linked document, the second is the ID of this
document. The databank provided may also be an alias for one or more
databanks.
And then the parser can override the content of the document by calling
set_document. See for an example the oxford parser.
If the documents do not contain unique ID fields, you can use the
next_sequence_nr call to generate unique numbers for a newly created databank
that can then be used to create IDs.
3. FastA extraction
If the parser is used for a databank that should have a FastaA file, it should
contain a to_fasta method. This method will receive four useful parameters
to use:
text
The entire document
db
The databank ID or mnemonic (e.g. 'sprot' for SwissProt)
id
The content of the id attribute, or the document number.
title
The title stored as attribute
The to_fasta function should return a FastA format containing an ID for each
sequence that looks like this:
>gnl|$db|$id|seqnr $title
The seqnr field is optional, it is only needed when multiple sequences can be
found in a single document (PDB e.g.).
This to_fasta function is called during indexing and can be called by the
web application.
|