pysisl 0.0.2

A python library for serialising and deserialising SISL (Simple Information Serialization Language)

pySISL

A python library for serialising and deserialising SISL (Simple Information Serialization Language). SISL is a simple structured text format designed for use in the NCSC Safely Importing Data Pattern. This library provides the ability to serialise and deserialise SISL as well as perform semantic verification of the SISL.

Hardware enforced syntatic verification may be carried out by the Oakdoor^TM family of data diodes.

Examples

Encoding basic Python dictionary hierarchies to SISL:

>>> import pysisl
>>> pysisl.dumps({"hello": "world"}) 
'{hello": !str "world"}'
>>> pysisl.dumps({"name": "helpful_name", "flag": False, "count": 3}) 
'{name: !str "helpful_name", flag: !bool "false", count: !int "3"}'

Decoding SISL to Python:

>>> import pysisl
>>> pysisl.loads('{name: !str "helpful_name", flag: !bool "false", count: !int "3"}') 
{'name': 'helpful_name', 'flag': False, 'count': 3}

Basic Usage

pysisl.dumps(dict)  

Serialise a python dictionary object into a SISL formatted str using this conversion table.

pysisl.loads(sisl, schema=None)  

Deserialise sisl str to a Python dictionary using this conversion table. Optionally, verify the sisl schema using a json schema.

Semantic Verification with a Schema

The jsonschema library is used to optionally verify the parsed sisl data structure. See JSON Schema for details on the json schema syntax. For example

Successful Parsing

>>> import pysisl
>>> my_schema = {
        "properties": {
            "name": {
                "type": "string"
            },
            "flag": {
                "type": "boolean" 
            },
            "count": {
                "type": "number"
            }
        }
    }
>>> decode_example = '{name: !str "helpful_name", flag: !bool "false", count: !int "3"}'
>>> pysisl.loads(decode_example, my_schema)
{'name': 'helpful_name', 'flag': False, 'count': 3}

Schema Verification Fails

>>> import pysisl
>>> my_schema = {
        "properties": {
            "name": {
                "type": "string"
            },
            "flag": {
                "type": "boolean" 
            },
            "count": {
                "type": "string"
            }
        }
    }
>>> decode_example = '{name: !str "helpful_name", flag: !bool "false", count: !int "3"}'
>>> pysisl.loads(decode_example, my_schema)
Traceback (most recent call last):
    File "/home/vagrant/pysisl/pysisl/sisl_decoder.py", line 31, in _verify_schema_if_required
        json_validator(flattened_sisl, schema=schema, format_checker=FormatChecker())
    File "/home/vagrant/pysisl/venv/lib64/python3.6/site-packages/jsonschema/validators.py", line 934, in validate
raise error
        jsonschema.exceptions.ValidationError: 3 is not of type 'string'

Failed validating 'type' in schema['properties']['count']:
        {'type': 'string'}

Conversion table

| Python | SISL |
| ------ | ---- |
| dict   | obj  |
| list   | list |
| str    | str  |
| int    | int  |
| float  | float|
| bool   | bool |
| None   | null |

Background

The NCSC Safely Importing Data Pattern, an architecture pattern describes a safe mechanism for handling structured data from an external untrusted source. We use a Transform - Verify approach taking our source data, transforming to an intermediate format, inspecting the intermediate format and then transforming back to the original format. SISL was designed to be a simple and easily inspectable intermediate format for just such an approach.

Oakdoor^TM products enable one- or two-way data transfers between segregated networks, letting organisations safely run services, such as file transfer, protocol exchanges, secure internet browsing and systems management. This is done using a combination of hardware enforced verification and software.

pySISL can form part of the transformation engine sub-system that enables cross-network communication that is compatible with the NCSC Safely importing data pattern. The pySISL encoder can be used to convert complex python dictionaries into valid SISL that is compatible with the diodes and the decoder will convert the SISL back into the same dictionaries without loss of data.

License

MIT licence

SISL Specification

For reference, this is ABNF for SISL.

grouping        = "{" *WSP ( name ":" 1*WSP "!" type 1*WSP value *WSP *( "," *WSP name ":" 1*WSP "!" type 1*WSP value *WSP ) / "" ) "}"

name            = ( "_" / ALPHA ) *( "_" / "-" / "." / ALPHA / DIGIT )

type            = ( "_" / ALPHA ) *( "_" / "-" / "." / ALPHA / DIGIT )

value           = ( DQUOTE *( PRINTABLE / ESCAPE) DQUOTE ) / grouping

ESCAPE          = "\" ( LCR / LCT / LCN / DQUOTE / "\" / (LCX 2HEX) / (LCU 4HEX) / (UCU 8HEX) )

HEX             = DIGIT / %x41-46 / %x61-66

WSP             = SP / HTAB / CR / LF

PRINTABLE       = %x20-21 / %x23-5B / %x5D-7E             ; Printable chars apart from '"' or '\'
ALPHA           = %x41-5A / %x61-7A                       ; A-Z / a-z
DIGIT           = %x30-39                                 ; 0-9
DQUOTE          = %x22                                    ; " (double-quote)
SP              = %x20                                    ; space
HTAB            = %x09                                    ; horizontal tab
CR              = %x0D                                    ; carriage return
LF              = %x0A                                    ; line feed
LCR             = %x72                                    ; lower case r
LCT             = %x74                                    ; lower case t
LCN             = %x6E                                    ; lower case n
LCX             = %x78                                    ; lower case x
LCU             = %x75                                    ; lower case u
UCU             = %x55                                    ; upper case uabnf/sisl spec publish

Getting Help

If you need help using the pySISL module, please contact Oakdoor^TM support at [e-mail].

Contributing to pySISL

All contributions, bug reports, bug fixes, documentation improvements, enhancements, and ideas are welcome.

If you notice a bug or would like to make an update to pySISL, please open an issue or raise a pull request.