tree 92ca19660614c182acb5766ae5cfe656e8a38683
parent e2c68b315fabf270b925c8c85aed5cfa907e77e8
author Matt Spinler <spinler@us.ibm.com> 1471285870 -0500
committer Patrick Williams <patrick@stwcx.xyz> 1473357857 +0000

Add ability to load inventory from file instead of hardcoding.

This commit allows inventory_items.py to load the inventory from
/usr/share/inventory/inventory.json instead of using the hardcoded
array in <System>.py.  A future commit will generate inventory.json
from the machine readable workbook XML for at least one system.

If the code can't load the json file, it will fall back to the
hardcoded array, which at this point we know is always present.

Even though pyinventorymgr will probably be rewritten in C++ at
some point, this series of commits illustrates how the OpenBMC
can utilize a data-driven design, and this component was chosen
as it takes a minimal amount of changes.

Change-Id: Ie749a068d6f47c5671677b21ff3ef5418739f161
Signed-off-by: Matt Spinler <spinler@us.ibm.com>