# Sequence diff hunks

`sequence-diff-hunks` · version 1.0.0 · Sequences & diffs · free, no key needed

Group a Myers sequence edit script into context hunks with 0-based original addresses.

**Use when you need to: sequence diff hunks · grouped sequence diff hunks · sequence context hunks.**

## Supported

- sequence diff hunks
- grouped sequence diff hunks
- sequence context hunks

## Not supported

- fuzzy matching
- patience diff
- histogram diff
- offset search
- RFC 6902 JSON Patch
- csv row diff

## Behavior

- left and right are sequences of JSON null, boolean, string, or IEEE-754 safe integer items.
- Hunks are built from the Myers SES via hunksFromSes. context defaults to 3, may be 0, and is at most 16.
- Changes separated by fewer than 2*context keep ops merge into one hunk.
- old_start and new_start are 0-based. When old_count is 0, old_start is the insertion index in [0, left.length].
- Hunk ops are {op, value} only. identical is true iff there are no change ops.
- For left ["A","B","C"] and right ["A","X","C"] with context 0 there is one hunk at old_start 1 replacing B with X.
- The string "__proto__" is an ordinary item. Unknown input fields are rejected.

## Input

- `left` (array of null or boolean or string or integer, required): max items 1024
- `right` (array of null or boolean or string or integer, required): max items 1024
- `context` (integer, optional): min 0; max 16

## Output

- `hunks` (array of any, required): max items 1024
- `identical` (boolean, required)

## Limits

- max sequence items: 1024
- max item string bytes: 4096
- max text bytes: 262144
- max context: 16
- max hunks: 1024
- max output bytes: 1048576

## Example

Request input:

```json
{
  "left": [
    "A",
    "B",
    "C"
  ],
  "right": [
    "A",
    "X",
    "C"
  ],
  "context": 0
}
```

Response:

```json
{
  "result": {
    "hunks": [
      {
        "old_start": 1,
        "old_count": 1,
        "new_start": 1,
        "new_count": 1,
        "ops": [
          {
            "op": "delete",
            "value": "B"
          },
          {
            "op": "insert",
            "value": "X"
          }
        ]
      }
    ],
    "identical": false
  }
}
```

## How to call it

### MCP

Connect `https://computefirst.net/mcp` ([setup](/docs#connect)), then call `execute` with:

```json
{
  "id": "sequence-diff-hunks",
  "version": "1.0.0",
  "input": {
    "left": [
      "A",
      "B",
      "C"
    ],
    "right": [
      "A",
      "X",
      "C"
    ],
    "context": 0
  }
}
```

### HTTP (no key)

```sh
curl -X POST https://computefirst.net/v1/tools/sequence-diff-hunks/versions/1.0.0/execute \
  -H "Content-Type: application/json" \
  -d '{"left":["A","B","C"],"right":["A","X","C"],"context":0}'
```

The machine-readable contract is at [/v1/tools/sequence-diff-hunks/versions/1.0.0](/v1/tools/sequence-diff-hunks/versions/1.0.0).

### CLI

```sh
node cli.mjs run sequence-diff-hunks 1.0.0 --input input.json --base-url https://computefirst.net
```

Get the client at [/clients/cli/](/clients/cli/).

## Related tools

- [Sequence apply hunks](/tools/sequence-apply-hunks): Apply disjoint original-address sequence hunks with exact context and no fuzz.
- [Sequence invert hunks](/tools/sequence-invert-hunks): Invert sequence hunks by swapping delete/insert and old/new counts.
- [Sequence apply edit script](/tools/sequence-apply-edit-script): Apply a keep/delete/insert edit script to a bounded sequence with exact value matching and no fuzz.
- [Sequence compose edit scripts](/tools/sequence-compose-edit-scripts): Compose two sequence edit scripts into one per-item keep/delete/insert script.
- [Sequence edit script](/tools/sequence-edit-script): Compute the Myers shortest edit script between two bounded sequences, with keep/delete/insert counts.
- [Sequence index map](/tools/sequence-index-map): Map each left-sequence index to a kept right index, or null if deleted, and list inserted right indices.