ccsds_ndm.kvn_handlers

Handler category sentinels and TypeHandler namedtuple for the KVN registry.

Each xsdata type is identified by its class name (str) and mapped to a TypeHandler namedtuple that classifies its locator, parser, and writer. Types absent from a schema registry implicitly use DEFAULT_HANDLER.

There is one registry dict per NDM schema version (e.g. ndmxml2.REGISTRY for ndmxml 2.0). The string class name can be resolved to the actual class at runtime via the schema’s __init__.py imports.

Handler categories

Locators (locate)

Classify a span of KvnLines as belonging to one instance of a type.

LOC_DEFAULT

Keyword-set matching via get_ccsds_kw_list. A new instance starts when a keyword belonging to this type’s keyword set appears, and ends at the first repetition of that keyword or at a block boundary. Works for every flat KEY = VALUE container (NdmHeader, OpmMetadata, StateVectorType, ManeuverParametersType, SpacecraftParametersType, …).

LOC_PACKED_STATE

Epoch-led packed lines (EPOCH X Y Z X_DOT Y_DOT Z_DOT [X_DDOT …]). Each logical record is one text line; a new instance starts on the next non-comment line. Used by StateVectorAccType (OEM data rows).

LOC_PACKED_ATTITUDE

Epoch-led packed lines whose column count and semantics depend on ATTITUDE_TYPE from the enclosing segment’s metadata. Used by AttitudeStateType (AEM data rows).

LOC_COVARIANCE

Starts on an EPOCH keyword line, followed by an optional COV_REF_FRAME keyword line, then 6 lower-triangular numeric rows (21 values total). Used by OemCovarianceMatrixType.

LOC_TDM_OBS

Each observation is KEY = EPOCH VALUE on one line inside DATA_START / DATA_STOP markers. Used by TrackingDataObservationType.

LOC_PACKED_LINES

Mixed block: KEY=VALUE header fields followed by a packed list[str] (trajLine / attLine / covLine). One instance starts when the anchor keyword (first declared uppercase field) appears; ends before the next occurrence of that same anchor. Used by OcmTrajStateType, OcmCovarianceMatrixType, AcmAttitudeStateType, AcmCovarianceMatrixType.

Parsers (parse)

Convert located KvnLines into an xsdata object.

PARSE_DEFAULT

Splits each line into [keyword, value, unit] and feeds the result to build_ndm_object from kvn_utils. Sub-containers are handled by recursing through the registry. Works for all KEY = VALUE types.

PARSE_PACKED_STATE

Splits one packed line on whitespace → positional columns for EPOCH, X, Y, Z, X_DOT, Y_DOT, Z_DOT (+ optional X_DDOT, Y_DDOT, Z_DDOT) → StateVectorAccType.

PARSE_PACKED_ATTITUDE

Splits one packed line on whitespace; derives the column template from ctx['meta'].attitude_type (and euler_rot_seq where needed), then populates the appropriate sub-object inside AttitudeStateType: one of quaternion_state, quaternion_derivative, quaternion_euler_rate, euler_angle, euler_angle_rate, spin, spin_nutation.

PARSE_COVARIANCE

Reads the EPOCH (and optional COV_REF_FRAME) keyword lines, then the 6 lower-triangular numeric rows → OemCovarianceMatrixType.

PARSE_TDM_OBS

Splits the KEY = EPOCH VALUE line into keyword, epoch, value → TrackingDataObservationType.

PARSE_ROTATION_ANGLE

The keyword is NOT fixed — it comes from the angle discriminator attribute (X_ANGLE / Y_ANGLE / Z_ANGLE). Reads up to 3 such lines → RotationAngleType.

PARSE_ROTATION_RATE

Analogous to PARSE_ROTATION_ANGLE for the rate discriminator (X_RATE / Y_RATE / Z_RATE) → RotationRateType.

PARSE_PACKED_LINES

Reads KEY=VALUE header fields via the default parser, then stores each remaining packed text line verbatim into the embedded list[str] field. Used by OcmTrajStateType, OcmCovarianceMatrixType, AcmAttitudeStateType, AcmCovarianceMatrixType.

Writers (write)

Serialise an xsdata object to a list of KvnLines.

WRITE_DEFAULT

Iterates dataclass fields in declaration order:

  • str / int / float / Enum leaf → one KvnLine (keyword from field metadata "name", value, unit from the units attribute on the value object if present).

  • list[str] (COMMENT) → one CommentLine per entry.

  • sub-container → recurse via registry.

Works for all KEY = VALUE containers.

WRITE_PACKED_STATE

Emits one line per StateVectorAccType: EPOCH X Y Z X_DOT Y_DOT Z_DOT [X_DDOT Y_DDOT Z_DDOT].

WRITE_PACKED_ATTITUDE

Emits one line per AttitudeStateType; column order derived from the populated sub-object and ctx['meta'].attitude_type.

WRITE_COVARIANCE

Emits EPOCH (and optional COV_REF_FRAME) keyword lines then the 6 lower-triangular numeric rows.

WRITE_TDM_OBS

Emits KEY = EPOCH VALUE for each TrackingDataObservationType.

WRITE_ROTATION_ANGLE

Emits up to 3 lines, keyword from rotation.angle (X_ANGLE / Y_ANGLE / Z_ANGLE) → RotationAngleType.

WRITE_ROTATION_RATE

Analogous to WRITE_ROTATION_ANGLE for the rate discriminator (X_RATE / Y_RATE / Z_RATE) → RotationRateType.

WRITE_PACKED_LINES

Emits the KEY=VALUE header fields first, then one text line per entry in the embedded packed list[str] field (trajLine / attLine / covLine). Used by OcmTrajStateType, OcmCovarianceMatrixType, AcmAttitudeStateType, and AcmCovarianceMatrixType.

Document-level dispatch (structural, not per-type)

DISPATCH_FLAT

OPM, OMM, APM, RDM — single segment, no META_START/STOP markers. Partition lines into header / metadata / data buckets by keyword-set membership, then call the registry for each bucket.

DISPATCH_SEGMENTED

OEM, AEM, TDM — one or more segments delimited by META_START / META_STOP / DATA_START / DATA_STOP (plus COVARIANCE_START / COVARIANCE_STOP for OEM). A state machine yields (meta_lines, data_lines[, cov_lines]) per segment before calling the registry.

DISPATCH_CDM

CDM — split on OBJECT_1 / OBJECT_2 keywords; handle the RELATIVE_METADATA_DATA section separately.

Classes

SchemaRegistry(*args, **kwargs)

Structural type for schema-specific KVN registries (ndmxml1, ndmxml2, …).

TypeHandler(locate, parse, write)

Handler classification for one xsdata type.
class SchemaRegistry(*args, **kwargs)[source]

Structural type for schema-specific KVN registries (ndmxml1, ndmxml2, …).

class TypeHandler(locate, parse, write)[source]

Handler classification for one xsdata type.

Each field holds a sentinel string from the constants above. All three slots must be explicitly filled. Types not present in a schema registry implicitly use DEFAULT_HANDLER.

Fields

locate:

Which locator strategy to use.

parse:

Which parser strategy to use.

write:

Which writer strategy to use.

Create new instance of TypeHandler(locate, parse, write)

locate: str

Alias for field number 0

parse: str

Alias for field number 1

write: str

Alias for field number 2

Parameters: