Introduction¶
Pre-release Disclaimer
This document is part of a pre-release version with work still in progress. And as such everything must be considered subject to change.
Scope¶
This document is the user manual for the ELT ICS Framework products OCM and DPM and covers:
Overview of the design and concepts
Interface documentation
The intended audience are instrument or instrument framework developers. Readers are assumed to be familiar with core ELT technologies:
CII MAL (request/reply and publish/subscribe)
FITS (Header/Data Units, keywords, extensions)
Acronyms¶
- CII
Core Integration Infrastructure
- DAQ
Data Acquisition
- DPM
Data Product Manager
In some contexts
DPM
may colloquially be referring to the applicationdaqDpmServer
.- DICD
Data Interface Control Document
- ELT
Extremely Large Telescope
- FITS
Flexible Image Transport System
- DHS
Data Handling Server where DPM deliver merged products to On-Line Archive System.
- OCM
Observation Coordination Manager
In some contexts
OCM
may colloquially be referring to the applicationdaqOcmServer
.- OCS
Observation Coordination System
- OLAS
On-Line Archive System
- OLDB
On-Line Database
- ICD
Interface Control Document (may also refer to MAL XML Interface document)
- ICS
Instrument Control System
- JSON
JavaScript Object Notation
- MAL
Middleware Abstraction Layer
- TBC
To be confirmed
- TBD
To be determined
Definitions¶
- Config Path
Term used in this document to refer to configuration paths resolved using the environment variable
$CFGPATH
. The path can either be absolute or relative. In case of relative path the file is resolved by testing the combination of each path in$CFGPATH
with the provided relative path.For example if
$CFGPATH=/tmp/a:/tmp/b
and Config Path is the filec
the tested paths are (in order):/tmp/a/c
/tmp/b/c
- Commentary Keyword
In this document a FITS Commentary Keyword refers to those keywords that are neither Value Keywords or ESO Keywords. Commentary Keywords are special because the same keyword name may occur multiple times in the same HDU, to e.g. continue a comment over multiple records.
Note
Strictly speaking the
HIERARCH
keyword is also a commentary keyword but is treated specially in applications that support the convention.Examples are:
COMMENT Commentary keyword names may occur multiple times in a header, it may also COMMENT be contextual, as in this case, where a comment is continued over multiple COMMENT records. HISTORY File modified by user 'USER' on host on 2021-09-03T01:28:26
- Data Acquisition
In this document it refers to the process of acquiring data as coordinated by OCM which results in one or more FITS files that contain the primary HDU and zero or more extensions such as binary tables, and or FITS keywords only. These are then merged together used to create a Data Product by DPM.
Refer to section Conceptual Model for a high level conceptual model of the process involved and Process Overview for a concrete and detailed overview.
- Data Product
In this document it refers to the FITS science data product that is produced by DPM using the acquired data (see Data Acquisition) and normally archived in the ESO Online Archive System.
- Data Product Specification
Specifies how to create a Data Product from input sources. This specification is created by daqDpmServer and used by daqDpmMerge.
- ESO Keyword
- ESO Hierarch Keyword
Refers to FITS keywords following the ESO HIERARCH keyword conventions RD6, i.e. keywords of the form:
HIERARCH ESO INS FILT1 ENC = 2 / Filter wheel absolute position [Enc]. HIERARCH ESO INS FILT1 ID = 'OUT' / Filter unique id.
The first token,
INS
in the example above, is referred to as the category.- Keyword Type
Refers to the different FITS standard[RD5] keyword record specifications which can be any of:
- Logical Keyword Name
The logical name of a FITS keyword depends on the type of keyword, but there are common traits: Trailing white spaces are insignificant and are not part of the Logical Keyword Name.
For FITS Value Keyword and Commentary Keyword the logical name is the white-space trimmed name component. The logical names for
NAXIS1 = 2048 / # of pixels in axis1 COMMENT This table was written by 'APPLICATION'
are
NAXIS1
andCOMMENT
respectively.For ESO Keyword the logical name is the part between
HIERARCH ESO
up to=
. For example the logical name isINS FILT1 ID
for the HIERARCH keyword:HIERARCH ESO INS FILT1 ID = 'OUT' / Filter unique id.
- metadaqif
Standard metadata source interface used by e.g. FCF and other components [RD3].
daqOcmServer
Name of the OCM main server executable.
daqOcmCtl
Name of the OCM command line client executable.
- recif
Standard primary data source interface implemented by DCSs [RD4].
- Value Keyword
A FITS Value Keyword are those keywords that have a value indicator in bytes 9 and 10 (c.f. RD5). Example keywords are:
SIMPLE = T / Standard FITS BITPIX = 8 / # of bits per pix value NAXIS = 0 / # of axes in data array
Style Conventions¶
JSON Data Structures¶
This manual documents a number of JSON data structures, their “schema”, using these common conventions. In some cases a more formal documentation in the form of JSON Schema is also provided.
Typically a concrete example is first provided:
{
"name": "example",
"optional": 3.0,
"nestedAnonymous": {
"nestedProperty": "value"
},
"listOfObjectsProperty": [
{
"type": "unionTypeA",
"property": "value"
},
{
"type": "unionTypeB",
"anotherProperty": "value"
}
]
}
This top level structure is then documented, sometimes with named sub-objects. The types of each
property uses Python Variable Annotation syntax. For
example a string would be str
, a JSON object would be object
, a list of strings would be
List[str]
, a union of two alternative named objects like above would use Union[type1,
type2]
, a list of union of named types would be List[Union[type1, type2]]
.
Unions are used when multiple types alternatives are allowed. The way it is typically used for is to
have multiple object alternatives. In this case the union variant is discriminated using a type
property with a fixed literal string value. The example in Listing 1 contain
a list of union type using this pattern.
Each property of an object, consisting of a <name>, <type> and <default> is then documented as follows:
<name>
(<type>) <default>Documentation for property with name “<name>” type “<type>” and optionally a default value <default>.
If <type> must have a specific value the literal value may be indicated instead.
If property is optional the <type> can specify this using Python annotation syntax:
Optional[<type>]
. If there is a default value <default> specifies that as[default DEFAULTVALUE]
. Lists are often optional and in this case it is simply indicated by a default empty list value:[default: []]
. When a property is optional and no value should be provided the property should simply be omitted from the object rather than specify a null value.If a property is an anonymous object with nested properties it will be documented inline with nested indentation (see example below).
Documenting the example structure above would look like:
Root object
name
(str)Name property.
optional
(Optional[float]) [default: 1.0]Optional property of float type with a default value of 1.0.
nestedAnonymous
(object)Example of how nested anonymous objects are documented.
nestedProperty
(str)A nested property of this anonymous object.
listOfObjectsProperty
(List[Union[UnionTypeA, UnionTypeB])Contains a list of either UnionTypeA or UnionTypeB types. See below for their structure.
UnionTypeA
Documents named object “UnionTypeA”.
type
(“unionTypeA”)Union discriminator and must have the literal value
"unionTypeA"
.property
(str)Example property of UnionTypeA.
UnionTypeB
Documents named object “UnionTypeB”.
type
(“unionTypeB”)Union discriminator and must have the literal value
"unionTypeB"
.anotherProperty
(str)Example property of UnionTypeB.