sed-flyt-neessi-rina

name: sed-flyt-neessi-rina description: >- Generates a file showing how the content of a SED is mapped between Neessi-dto, ACL-SED and RINA-SED. Use this when asked to generate a SED flow mapping, sedFlyt, or field mapping for any SED type (e.g. H120, F001, H065). allowed-tools: github-mcp-server-get_file_contents license: "Internal / NAV"

sed-flyt-neessi-rina — SED Flow Mapping Generator

Generates a tab-aligned mapping file for a given SED type showing the relationship between:

• indeks — field position from dokumentasjon (Rina)
• neessi-felt — eux-neessi field path (Neessi-dto)
• aclSedSti — ACL SED path from eux-acl-generator (Acl-sed)
• Rina path — full Rina path (Rina-sed)

Output file: doc/<SED>_sedFlyt_NeessiRina.txt in the eux-architecture repo.

Input

The user provides any SED type, e.g. H065, H120, F001, U001, X008.

Source data locations

All source files are fetched from GitHub (organization: navikt). Use the github-mcp-server-get_file_contents tool to retrieve them.

From navikt/eux-acl-generator:

• nyAclMal_basis.txt: src/main/resources/endringer/cdmAclMaler/v44/<letter>/<SED>/nyAclMal_basis.txt where <letter> is the first letter(s) of the SED type (H, F, X, etc.)
• dokumentasjon JSON: src/main/resources/resultater/dokumentasjon/v44/<letter-lowercase>/<SED>_dokumentasjon.json

From navikt/eux-neessi (these are optional — many SED types have neither):

• DTO: src/main/java/no/nav/eux/neessi/restapi/sed/dto/<SED>Dto.java
• Service: src/main/java/no/nav/eux/neessi/service/sed/<SED>Service.java
• Model class: src/main/java/no/nav/eux/neessi/model/sed/<letter>/<SED>.java

Steps

1. Locate source files

Fetch the following from GitHub using github-mcp-server-get_file_contents (owner: navikt):

# From repo: eux-acl-generator
acl_file = src/main/resources/endringer/cdmAclMaler/v44/<LETTER>/<SED>/nyAclMal_basis.txt
dok_file = src/main/resources/resultater/dokumentasjon/v44/<letter>/<SED>_dokumentasjon.json

# From repo: eux-neessi (all optional — may not exist)
dto_file = src/main/java/no/nav/eux/neessi/restapi/sed/dto/<SED>Dto.java
svc_file = src/main/java/no/nav/eux/neessi/service/sed/<SED>Service.java
model_file = src/main/java/no/nav/eux/neessi/model/sed/<letter>/<SED>.java

Verify that acl_file and dok_file exist (the GitHub tool returns an error for missing files). If not, abort with an error message. The dto_file, svc_file, and model_file are optional — many SED types have neither.

Priority: If dto_file exists, use the DTO approach (step 4a). Otherwise, if model_file exists, use the model class approach (step 4b, see model-classes.md). Otherwise, if the SED is one of F003, F026, F027, use the FamilieSed class approach (step 4c, see familiesed-mapping.md). If none of the above apply, produce a 3-column output (indeks, aclSedSti, ACL path).

2. Parse nyAclMal_basis.txt

Each line has one of these formats:

• H120.Some.Path: $acl.sed.sti — ACL path with aclSedSti mapping
• H120.Some.Path: : EnumTypeName — ACL path with enum type (no aclSedSti)
• H120.Some.Path — ACL path without mapping (structural/intermediate node)

Parse into a dict: { ACL_path: aclSedSti_or_empty }.

3. Parse dokumentasjon JSON

The JSON is keyed by ACL path. Each entry has:

• indeks — hierarchical field position (e.g. "1.1.7.1.4.2.")
• aclSedSti — ACL SED path (may differ from nyAclMal_basis — dokumentasjon uses .landkode where nyAclMal_basis uses .land)
• required — boolean
• ledetekst, verdiType, cdmType — metadata (not used in output)

Use this to get the indeks for each ACL path. If nyAclMal_basis has no aclSedSti for a path, fall back to dokumentasjon's aclSedSti.

4a. Build DTO mapping (if DTO exists)

If <SED>Dto.java exists:

1. Parse the record hierarchy to extract all leaf field paths (e.g. H120Dto.bruker.personInfo.etternavn).
2. Read <SED>Service.java to understand the mapping between aclSedSti and DTO fields. The Service has tilDto() and tilEuxRinaApiDto() methods that show the field-by-field mapping.
3. Build a dict: { aclSedSti: DTO_field_path }.

Key mapping gotchas:

• nyAclMal_basis uses .land but DTOs use .landkode — handle both variants.
• [x] in aclSedSti corresponds to [] in DTO paths.
• Some DTO fields map to multiple ACL paths (e.g. dekningKostnader maps to both 4.1.2 and 4.2.1).
• Some DTO fields have no ACL mapping (metadata, PDL enrichment, backend-set fields).

4b. Build model class mapping (if model class exists but no DTO)

See model-classes.md for the full model class hierarchy, field tracing rules, and mapping gotchas.

4c. Build FamilieSed mapping (for F003, F026, F027)

See familiesed-mapping.md for FamilieSed fields, supporting classes, and mapping rules.

5. Handle .land / .landkode variants

For classes that have BOTH .land and .landkode fields (Statsborgerskap, Adresse, PinMangler.Foedested):

• Show both variants as separate rows.
• The .landkode sibling row has empty indeks (do not repeat).
• Fill in the .landkode ACL path in column 3 if it exists in dokumentasjon.

For Pin (which only has .landkode): show a sibling row with the .landkode ACL path if it exists in dokumentasjon.

6. Assemble rows

For each entry in nyAclMal_basis (and any extra entries from dokumentasjon with aclSedSti):

(indeks, neessi_field_path_or_empty, aclSedSti, acl_path)

Sort by indeks using natural numeric sort on the hierarchical parts:

• Split indeks by ., parse each part as integer
• Sort as tuple: (1, 1, 7, 1, 4, 2) for "1.1.7.1.4.2."

After sorting, insert .landkode sibling rows immediately after the corresponding .land row.

7. Add unmapped neessi fields

Find fields in the DTO/model class not covered by any ACL row. Add them at the bottom with:

• Empty indeks (column 1)
• Field path (column 2)
• Explanation in parentheses (column 3), e.g. (metadata, ikkje i ACL), (beriking frå PDL, ikkje i ACL), (sett av backend, ikkje i ACL), (ikkje i <SED> ACL), (les frå <path>, ikkje i <SED> ACL)
• Empty column 4

8. Format output

Columns (tab-separated) — always 4 columns when a DTO or model class exists:

If neither DTO nor model class exists, use 3 columns (indeks, aclSedSti, ACL path).

Tab alignment: Use the format-output.py script in this skill's directory. It reads tab-separated rows from stdin and applies correct tab alignment:

• tab_size = 8
• For each column, compute minimum fixed width: target = ((max_width // 8) + 1) * 8
• Tabs needed per cell: max(1, ceil((target - cell_length) / 8))

Section separators: Insert a blank line between rows where the first digit of the indeks changes (i.e. between sections 1, 2, 3, etc.).

9. Write output

Write to doc/<SED>_sedFlyt_NeessiRina.txt in the eux-architecture repo (current working directory).

Report:

• Number of ACL rows
• Number of rows with neessi mapping
• Number of .landkode sibling rows added
• Number of unmapped neessi fields (if applicable)
• Total lines written