Commands¶
Glot has a number of commands that allow it to act as a complete
client for Glossia. These should be included as described in
Usage, bearing in mind that global arguments, preceding the
command, may or may not modify behaviour
(e.g. --to
is relevant only for commands retrieving files
from the Glossia server).
The most commonly used commands are likely to be:
glot table
glot results -i UUID
(successful simulations)glot diagnostic -i UUID
(failed simulations)glot diagnostic -d UUID
(on remote host)TransferUUID-results.tgz
to local machineglot inspect UUID-results.tgz
(local machine)
In general, UUIDs may be given only as the first few characters, provided that this specifies them uniquely on the server.
Search¶
Search UUID or UUIDs matching a given prefix. Prints a table showing most recent status of matching simulations on the Glossia server.
glot search [--limit LIMIT] [--server-limit SERVERLIMIT]
[--sort SORT] [GUID]
Argument / Option | Description |
---|---|
GUID | (prefix of or) GUID to search for on server |
–limit LIMIT | maximum number of entries to display (default: 10) |
–server-limit SERVERLIMIT | maximum number of entries for the server to retrieve from the DB (NB: pre-sorting) (default: 1000) |
–sort SORT | order returned entries by SORT, which may be either ‘timestamp’ or ‘guid’ (default: timestamp) |
Table¶
Show a table of simulations. Slightly prettier, more intuitive alternative to using
glot search
with no arguments.
glot table [--limit LIMIT] [--server-limit SERVERLIMIT] [--sort SORT]
Argument / Option | Description |
---|---|
–limit LIMIT | maximum number of entries to display (default: 10) |
–server-limit SERVERLIMIT | maximum number of entries for the server to retrieve from the DB (NB: pre-sorting) (default: 1000) |
–sort SORT | order returned entries by SORT, which may be either ‘timestamp’ or ‘guid’ (default: timestamp) |
Results¶
Retrieve results of a simulation. Note that this does not retrieve run-time output, which Glossia
has no access to, only output that has been returned from simulation container. For manual
inspection, this is most useful
in conjunction with the -i
or -d
flags, which bundle diagnostic output.
glot results [--target TARGET] ([--include-diagnostic/-d] | [--inspect-diagnostic/-i]) GUID
Argument / Option | Description |
---|---|
GUID | (prefix of or) GUID to cancel on server. Must be unique |
–target TARGET | server for Glossia to push output bundles to using HTTP POST commands (on port 18103). By default, Glot will start an HTTP server and Glossia will look for it on the network gateway IP |
–include-diagnostic | combine this command with the diagnostic command to
grab both bundles of data |
–inspect-diagnostic | combine this command with the diagnostic and
inspect commands to produce a ready-to-run local
simulation |
Cancel¶
Interrupt a running simulation. Will raise an E_CANCELLED
error on Glossia
(more information).
glot cancel GUID
Argument / Option | Description |
---|---|
GUID | (prefix of or) GUID to cancel on server. Must be unique |
Status¶
Retrieve status information about a (current or ended) simulation. Particularly useful for
retrieving status or error messages that span multiple lines and cannot be fully displayed
by glot search
or glot table
.
glot status GUID
Argument / Option | Description |
---|---|
GUID | (prefix of or) GUID to cancel on server. Must be unique |
Diagnostic¶
Retrieve diagnostic output of a (successful or) failed simulation. This includes log output to STDERR/STDOUT, any input and any settings files supplied by Glossia to the container. Glossia stores the UUID of container image used for simulation within the diagnostic data.
glot diagnostic [--target TARGET] [--inspect/-i] GUID
Argument / Option | Description |
---|---|
GUID | (prefix of or) GUID to cancel on server. Must be unique |
–target TARGET | server for Glossia to push output bundles to using HTTP POST commands (on port 18103). By default, Glot will start an HTTP server and Glossia will look for it on the network gateway IP |
–inspect | combine this command with the inspect command to
achieve a ready-to-run simulation in the ./UUID/
directory |
Launch¶
Launch a new simulation.
glot launch [--tmp-directory TMPDIR] [--input/-i INPUT1 -i INPUT2 ...]
[--definition/-d DEF1 -d DEF2 ...] GSSAXML
Argument / Option | Description |
---|---|
GSSAXML | A file containing GSSA-XML defining the simulation (this is original.xml in diagnostic output) |
–tmp-directory TMPDIR | directory to use for exchanging files with Glossia (default: /tmp) |
–input INPUTN | (with multiplicity) files that are referenced with the GSSA-XML and are expected in the input/ subfolder of the simulation directory |
–definition DEFN | (with multiplicity) files that constitute the internal configuration. This corresponds to the content of the numerical model in the Go-Smart CDM framework. For instance, a Jinja2 SIF file for Goosefoot or a Python script and custom modules for a FEniCS simulation. |
Inspect¶
Take a diagnostic bundle and prepare it to be run as a new simulation locally (without Glossia, but using a simulation container). This allows precise reproduction of the remote behaviour (with certain standard Docker caveats, regarding kernel versions, etc.). To do so, it untars the diagnostic bundle, and clones a remote repository containing tools to execute the simulation. For more information on executing standalone simulations, see the individual mode repositories.
- Modes:
- fenics (not yet incorporated into glot)
glot inspect [--mode MODE] ARCHIVE
Argument / Option | Description |
---|---|
ARCHIVE | bundle retrieved by glot diagnostic , as TGZ |
–mode | type of setup that should be used (corresponds to a tool repository to clone into the diagnostic data). At present the only valid mode is ‘goosefoot’ (default) |