pypdx

A Simple parser for PDX (Product Data eXchange) XML files

Introduction

From the Wikipedia article on PDX: "the PDX (Product Data eXchange) standard for manufacturing is a multi-part standard, represented by the IPC 2570 series of specifications."

As the name implies, it is a standard for exchanging product definition between companies or organizations, and can include bill of material (BOM), approved manufacturer list, drawings, documents, etc.; pretty much anything can be included if desired.

In simple terms, a *.pdx file (usually exported from Agile/Oracle) is really just a ZIP file that contains all the files ("attachments") associated with the product (assembly), plus a special XML file called pdx.xml. This XML file contains the particulars of the various Items and their properties/attributes, the relationship between the Items, which forms the bill-of-materials (BOM), and also information about the various files that are inside the PDX/ZIP file.

The DTD of this XML file (identified as "DTD 2571 200111") can be found on the IPC website. Free PDX viewers are available, one of the most popular being PDXViewer from PDXplorer.

Since the *.pdx file is simply a ZIP file, it is easy to extract all the attachments (documents, schematics, drawings, etc.) from the ZIP file, and there are several ways one can view XML files. However, trying to make sense of the assembly from a generic XML viewer is not really feasible, and although there are specialized free (and non-free) viewers, there are times when you might want to extract the data for your own use. To this end I have written a simple PDX XML file parser, pypdx (written in Python), as presented here.

The pypdx program can be used as a Python module, but also as a command-line stand-alone program that is more or less ready to use. Be warned that it does. It does not implement all the elements defined in the PDX standard, but only the parts that happen to be of interest to me at the moment, so don't be surprised if it does not have the features you are looking for. What it does extract are the Items, the ApprovedManufacturerList, the BillOfMaterial, and the Attachments. The main purpose of the program is to extract this data from the XML file, and organize them by saving them into a relational database: in this case either a SQLite3 database or a PostgreSQL database. What you do after the data is stuffed into a relational database is up to you!

Installation and Usage

To install the module and program, run

pip install pypdx

This should create an executable pypdx. The usage is as follows:

USAGE: pypdx pdx-file.xml dsn [dump [remove_all_first]]
     - pdx-file.xml: this is the PDX XML file
     - dsn: can be a SQLite3 file (the program will create one if it does not exist; use the extension .sqlite3
          : or it can be the dsn connection string for a PostgreSQL database
          : if dsn is 'pg', the default database 'pdx' on localhost (port 5432) and username 'pdxuser' will be used
     - dump: 1 or 0; to dump to a JSON file pdx-dump.json (optional)
     - remove_all: 1 or 0; remove all records from the tables first (optional);
     -           : WARNING: this will delete *all* existing parts, BOM, etc., records from the database

Examples: for a SQLite3 database:

pypdx data/pdx-example.xml testout.sqlite3 1 1

for a postgreSQL database:

pypdx data/pdx-example.xml "dbname='pdx' host='localhost' user='pdxuser' password='billofmaterials' port=5432" 1 1

A sample PDX XML file (data/pdx-example.xml) is included in the distribution. This is obviously not for a real product.

To use pypdx as a module, do something like this:

import pypdx
dsn = 'testout.sqlite3'
xmlfile = 'data/pdx.xml'
mypdx = pypdx.PDX(xmlfile, dsn, debug=True)
# should return 'ok', otherwise you get the error message
status = mypdx.removeall()
status = mypdx.fillparts()

mypdx.removeall() removes all old records from the database tables, mypdx.fillparts() then fills the database table with new records from the XML file.

Database and Tables

The data is saved into the following tables:

partsmaster is the main table that stores the Items; each Item is uniquely identified by a itemUniqueIdentifier
bom is the table the stores the BOM; each BOM record is essentially a "link" that points from the parent Item (identified by its itemUniqueIdentifier), to the target Item (also identified by its itemUniqueIdentifier). (Viewed as a directed graph, the records in the partsmaster are the nodes, and the records of the bom are the edges of the graph.)
attachments is the table that stores information on the attachment files, which should be in the ZIP file if the isFileIn field is TRUE. There can be multiple attachments to each Item, and these are linked to the Item through the Item's itemUniqueIdentifier.
approvedmfg is the table that stores the information on the approved manufacturers for each Item. There can also be mutiple approved manufacturers for each Item.

The definition of (and relations between) these tables are laid out in the SQL files in the data/ directory. A sample program (main.py) in the source code distribution illustrates the usage of the module (this is used to form the pypdx program mentioned above). You can also dump the contents of the PDX file into a JSON file (with the PDX.dump(filename) function). However this merely mirrors the structure and contents of the XML file; it may not be particularly useful unless you process the JSON file/object further on your own.

As mentioned above, the program allows you to extract the data into a SQLite3 database or a PostgreSQL database. The former is less trouble to set up, as it is file-based. The program will in fact create the SQLite3 database file for you (as well as create the tables).

For using this in a PostgreSQL database, the program pypdx will create the tables for you if they do not already exist, but it assumes that the database called pdx already exists and is running on localhost (at port 5432). You can create the database with the commands:

% psql template1
....
template1=# create database pdx;
template1=# \q

or you can modify the dsn specifications in the example program to suit your needs. It should be relatively simple to modify the code to use a MySQL database, but I have not tried this.

The program depends on a few Python modules (specified in the requirement.txt file), including the SQLite3 driver (sqlite3) and the PostgreSQL driver (psycopg2). Run

% sudo pip install -r requirement.txt

to install the modules. If you do not care for the PostgreSQL database, you should still be able to use the program without installing the psycopg2 module, since it is not imported unless you specify the PostgreSQL database option.

A Simpler Viewer

For a more friendly user interface, and for viewing the contents of the PDX XML files, have a look at my simple web app pdxdisplay (written as a Flask application). The pdxdisplay program has also been packaged as a docker container, together with a PostgreSQL database.

Closing Remarks

I have only seen a very small number of PDX files, and there does not seem to be any sample PDX files that you can download from the Internet (likely because the only PDX files available contain proprietary manufacturing information!). Naturally the testing of this program has been very limited. While I believe the implementation to be correct (if incomplete), there is always the possibility of bugs. So use at your own risk; you have been warned!

(Last Revised 2018-02-03)