&pyb; is in fact a simple set of classes and functions written in python, and that provides with a simple and homogenous access to bibliographic databases. Therefore, it is possible to write python scripts that make use of these specialized functions. The graphical interface, &pyc;, is itself a simple script on top of &pyb;. To execute a script written for &pyb;, simply run > pybliographer myscript.py You can alternatively start your script by #!/path/to/pybliographer ...rest of your script... and make it executable. Some scripts are provided with pybliographer, both as example and as useful tools. They are quite short and should be readable with basical knowledge of python. This tool takes a list of files or directory, and check if they are valid (syntax, no entries with the same key,...) It is possible to use its output directly in an emacs compile buffer, in order to jump directly to the encountered errors. To do so, type M-x compile, then the command pybcheck yourfiles, and use the middle button of the mouse to jump into the faulty file. This tool extracts the citations made in a LaTeX document and generates a BibTeX file containing them. Usually, one stores its bibliographies in one or several large BibTeX files, and lets bibtex extract the entries used in a LaTeX document. But it is sometime convenient to create a self-contained package (for example to share it in native form with somebody else, or to store it), with a minimalistic BibTeX file holding exactly the entries used in LaTeX. This tool does exactly that: it reads a LaTeX .aux file, and extract from the specified BibTeX databases the corresponding entries. This tool converts from one bibliographic format to another. The general syntax is pretty simple. To convert from Refer to BibTeX for example, just run: > pybconvert refer..bibtex toto.refer toto.bib This script generates a bibliography according to a bibliographic style, and outputs it in a specific format (like HTML, LaTeX,...) as it should appear in a document. The general form of the command is > pybformat [options] database... This command will use the format called alpha to create a bibliography in LaTeX. Several options are available to create these documents: --style=... or -s ...: specify a bibliography style. This can be a full path to an existing XML file, or a name which will be searched in the standard places. Default is Alpha. --format=... or -f ...: specify an output format (like HTML, LaTeX,...). Default is Text. --output=... or -o ...: specify an output filename. STDOUT is the default. --header=... or -H ...: defines a file that will be prepended to the output file. --footer=... or -F ...: defines a file that will be appended to the output file. --list=output or -l output: lists the available output formats To start writing your own script, you can read what follows, and then have a look at the existing scripts. Trying to adapt them to fit your personal needs can be a good way of testing what you read. This section describes some basic classes and ideas that are useful to understand how &pyb; works. More information can be found in . In the following, all the modules that will be refered to belong to the Pyblio domain. So, to access the members of the Open module, you'll have to write at the beginning of your script something like from Pyblio import Open The Base module contains some of the most basical classes used in the application: Base.Entry represents a specific bibliographic entry, with all its fields. It behaves like a hash table which returns the content of a field given its name Base.DataBase is the class from which every database type inherits. It behaves like a hash table that returns a Base.Entry given a Base.Key Key.Key is the object that uniquely identifies an entry. This object must be unique over the whole application, and is composed of a database part and an entry part. An Iterator is an object that provides a way to access a sequence of items in order. These iterators are used extensively in &pyb;, because they hide the underlying access mechanism, and provide the same access on any database. They are also perfectly suited for implementing transparent filtering and sorting of entries: the Selection.Selection class for example takes an iterator (on a database for example) and return a new one which will only iterate on a subset of the entries, according to a search criterion.