The Data Management Framework (DMF) is used to manage all the data needed by the IDAES framework, including flowsheets, models, and results. It stores metadata and data in persistent storage. It does not require that the user run a server or connect to a remote service. The DMF can be accessed through its API or command-line interfaces.
The DMF is designed to allow multiple separate threads of work. These are
workspaces. Inside a given workspace, all the information is
represented by containers called
resources. A resource describes some
data in the system in a standard way, so it can be searched and manipulated
by the rest of the IDAES framework.
Resources can be connected to each other with
relations such as
“derived”, “contains”, “uses”, and “version”.
Below is an illustration of these components.
The DMF is configured with an optional global configuration file and a
required per-workspace configuration file. By default the global file is
looked for as
.dmf in the user’s home directory. Its main function at the
moment is to set the default workspace directory with the
keyword. For example:
# global DMF configuration
The per-workspace configuration has more options. See the documentation
Workspace class for details.
The configuration file is in YAML (or JSON) format. Here is an example file, with some
description in comments:
settings: # Global settings
workspace: /home/myuser/ws # Path to current workspace
workspace: # Per-workspace settings
location: /home/myuser/ws # Path to this workspace
name: myws # Name of this workspace
description: my workspace # Description (if any) of this workspace
created: 2019-04-09 12:55:05 # Date workspace was created
modified: 2019-04-09 12:55:05 # Date workspace was modified
files: # Basic information about data files
count: 3 # How many files
total_size: 1.3 MB # Total size of the files
html_documentation_paths: # List of paths for HTML documentation
logging: # Logging configuration
idaes.core.dmf: # Name of the logger
level: DEBUG # Log level (Python logging constant)
output: /tmp/debug.log # File path or "_stdout_" or "_stderr_"
This configuration file is used whether you use the DMF from the command-line,
Jupyter notebook, or in a Python program. For details see the
DMF package documentation.
In the Jupyter Notebook, there are some “magics” defined that make initializing the DMF pretty easy. For example:
from idaes.core.dmf import magics
%dmf init path/to/workspace
The code above loads the “%dmf” line magic in the first line, then uses it to initialize the DMF with the workspace at “path/to/workspace”.
From there, other “line magics” will operate in the context of that DMF workspace.
%dmf help- Provide help on IDAES objects and classes. See dmf-help.
%dmf info- Provide information about DMF current state for whatever ‘topics’ are provided
%dmf list- List resources in the current workspace
%dmf workspaces- List DMF workspaces; you can do this before %dmf init
The IDAES Python interfaces are documented with Sphinx. This includes
automatic translation of the comments and structure of the code into
formatted and hyperlinked HTML pages. The
%dmf help command lets you easily
pull up this documentation for an IDAES module, class, or
object. Below are a couple of examples:
# Initialize the DMF first
from idaes.core.dmf import magics
%dmf init path/to/workspace create
# Get help on a module (imported)
from idaes.core import control_volume1d
%dmf help control_volume1d
# Get help on a module (by name, no import)
%dmf help idaes.core.control_volume0d
# Get help on a class
from idaes.core.control_volume1d import ControlVolume1DBlock
%dmf help ControlVolume1DBlock
# Get help on a class (by name, no import)
%dmf help idaes.core.control_volume1d.ControlVolume1DBlock
# Get help on an object (will show help for the object's class)
# This will end up showing the same help as the previous two examples
obj = control_volume1d.ControlVolume1DBlock()
%dmf help obj
The help pages will open in a new window. The location of the built
documentation that they use is configured in the per-workspace DMF
configuration under the
htmldocs keyword (a default value is filled in
when the DMF is first initialized).
idaes.core.dmf package documentation that is generated
automatically from the source code.