Spaces:
Paused
Paused
File size: 12,231 Bytes
80f50ca | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 | """
Codebook storage (universal).
SQLite-backed CRUD over `codes` and `annotation_codes` in
`<task_dir>/project.sqlite` via the universal persistence layer. No
business rules live here (no cycle checks, no permissions, no cache
invalidation) — that is the service layer's job. This module only
persists rows.
A *code* is a (possibly nested) label in a project's codebook. An
*annotation_code* links a stored annotation to a code, optionally with a
time span (`started_at`/`ended_at`) for temporal / agentic-trace coding.
Universal: usable in standard annotation, solo mode, and QDA mode.
Design notes:
- `parent_id` is TEXT NOT NULL DEFAULT '' where '' means "root". Using a
sentinel instead of NULL lets `UNIQUE(project, parent_id, name)`
actually prevent duplicate sibling names at the root too (SQLite
treats NULLs as distinct in UNIQUE constraints).
- No SQL foreign keys (consistent with the memos store): the service
layer enforces parent existence, cycle-freedom, and recursive delete.
"""
from __future__ import annotations
import time
import uuid
from typing import Any, Dict, List, Optional
from potato.persistence import Migration, get_db, register_migration
ROOT = "" # sentinel parent_id for top-level codes
_CODEBOOK_MIGRATION = Migration(
name="0001_codebook",
sql="""
CREATE TABLE IF NOT EXISTS codes (
id TEXT PRIMARY KEY,
project TEXT NOT NULL,
name TEXT NOT NULL,
color TEXT,
parent_id TEXT NOT NULL DEFAULT '',
sort_order INTEGER NOT NULL DEFAULT 0,
created_by TEXT NOT NULL,
created_at REAL NOT NULL,
updated_at REAL NOT NULL,
UNIQUE (project, parent_id, name)
);
CREATE INDEX IF NOT EXISTS idx_codes_project ON codes (project);
CREATE INDEX IF NOT EXISTS idx_codes_parent
ON codes (project, parent_id);
CREATE TABLE IF NOT EXISTS annotation_codes (
annotation_id TEXT NOT NULL,
code_id TEXT NOT NULL,
project TEXT NOT NULL,
created_by TEXT NOT NULL,
started_at REAL,
ended_at REAL,
PRIMARY KEY (annotation_id, code_id)
);
CREATE INDEX IF NOT EXISTS idx_anncodes_code
ON annotation_codes (project, code_id);
CREATE INDEX IF NOT EXISTS idx_anncodes_ann
ON annotation_codes (annotation_id);
""",
)
register_migration(_CODEBOOK_MIGRATION)
def _db(task_dir: str):
"""Connection guaranteeing the codebook migration is registered.
register_migration is idempotent, so this is a no-op normally; it
makes the store robust if a test helper (clear_migrations) wiped the
process-global registry before this task_dir's first get_db().
"""
register_migration(_CODEBOOK_MIGRATION)
return get_db(task_dir)
def _ensure_temporal_schema() -> None:
"""Guarantee the Phase 2 (C) append-only columns/tables exist before
any link read/write that depends on `invalidated_at`. Lazy import
avoids a module-load cycle (changelog imports this module). The
0003 migration is additive (nullable cols + new tables), so this is
safe even for callers that only registered 0001."""
from potato.codebook.changelog import _CHANGE_MIGRATION
from potato.codebook.revision import (
_REVISION_MIGRATION, _CODES_REV_MIGRATION)
register_migration(_REVISION_MIGRATION)
register_migration(_CODES_REV_MIGRATION)
register_migration(_CHANGE_MIGRATION)
# ---- codes ---------------------------------------------------------------
def insert_code(
task_dir: str,
*,
project: str,
name: str,
created_by: str,
color: Optional[str] = None,
parent_id: str = ROOT,
sort_order: int = 0,
code_id: Optional[str] = None,
created_revision: int = 0,
) -> Dict[str, Any]:
"""Insert one code row and return it. `code_id` lets the init CLI
supply a deterministic id; otherwise a random uuid4 is used.
`created_revision` records the codebook revision the code first
appeared in (for provenance / the review worklist)."""
cid = code_id or uuid.uuid4().hex
now = time.time()
conn = _db(task_dir)
conn.execute(
"""INSERT INTO codes
(id, project, name, color, parent_id, sort_order,
created_by, created_at, updated_at, created_revision)
VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?)""",
(cid, project, name, color, parent_id, sort_order,
created_by, now, now, created_revision),
)
conn.commit()
return get_code(task_dir, cid)
def get_code(task_dir: str, code_id: str) -> Optional[Dict[str, Any]]:
row = _db(task_dir).execute(
"SELECT * FROM codes WHERE id = ?", (code_id,)
).fetchone()
return dict(row) if row else None
def find_code(
task_dir: str, project: str, parent_id: str, name: str
) -> Optional[Dict[str, Any]]:
row = _db(task_dir).execute(
"""SELECT * FROM codes
WHERE project = ? AND parent_id = ? AND name = ?""",
(project, parent_id, name),
).fetchone()
return dict(row) if row else None
def list_codes(task_dir: str, project: str) -> List[Dict[str, Any]]:
rows = _db(task_dir).execute(
"""SELECT * FROM codes WHERE project = ?
ORDER BY parent_id ASC, sort_order ASC, name ASC""",
(project,),
).fetchall()
return [dict(r) for r in rows]
def children_of(
task_dir: str, project: str, parent_id: str
) -> List[Dict[str, Any]]:
rows = _db(task_dir).execute(
"""SELECT * FROM codes WHERE project = ? AND parent_id = ?
ORDER BY sort_order ASC, name ASC""",
(project, parent_id),
).fetchall()
return [dict(r) for r in rows]
def update_code(
task_dir: str,
code_id: str,
*,
name: Optional[str] = None,
color: Optional[str] = None,
parent_id: Optional[str] = None,
sort_order: Optional[int] = None,
) -> Optional[Dict[str, Any]]:
sets, params = [], []
if name is not None:
sets.append("name = ?"); params.append(name)
if color is not None:
sets.append("color = ?"); params.append(color)
if parent_id is not None:
sets.append("parent_id = ?"); params.append(parent_id)
if sort_order is not None:
sets.append("sort_order = ?"); params.append(sort_order)
if not sets:
return get_code(task_dir, code_id)
sets.append("updated_at = ?"); params.append(time.time())
params.append(code_id)
conn = _db(task_dir)
conn.execute(f"UPDATE codes SET {', '.join(sets)} WHERE id = ?", params)
conn.commit()
return get_code(task_dir, code_id)
def delete_codes(task_dir: str, code_ids: List[str]) -> int:
"""Delete the given codes and their annotation_codes links. The
service computes the full subtree; this just executes the delete."""
if not code_ids:
return 0
qs = ",".join("?" * len(code_ids))
conn = _db(task_dir)
conn.execute(
f"DELETE FROM annotation_codes WHERE code_id IN ({qs})", code_ids)
cur = conn.execute(
f"DELETE FROM codes WHERE id IN ({qs})", code_ids)
conn.commit()
return cur.rowcount
# ---- annotation_codes ----------------------------------------------------
def link_annotation(
task_dir: str,
*,
project: str,
annotation_id: str,
code_id: str,
created_by: str,
started_at: Optional[float] = None,
ended_at: Optional[float] = None,
) -> None:
conn = _db(task_dir)
conn.execute(
"""INSERT OR REPLACE INTO annotation_codes
(annotation_id, code_id, project, created_by,
started_at, ended_at)
VALUES (?, ?, ?, ?, ?, ?)""",
(annotation_id, code_id, project, created_by,
started_at, ended_at),
)
conn.commit()
def unlink_annotation(
task_dir: str, annotation_id: str, code_id: str
) -> bool:
conn = _db(task_dir)
cur = conn.execute(
"""DELETE FROM annotation_codes
WHERE annotation_id = ? AND code_id = ?""",
(annotation_id, code_id),
)
conn.commit()
return cur.rowcount > 0
def codes_for_annotation(
task_dir: str, annotation_id: str
) -> List[Dict[str, Any]]:
# THE single load-bearing temporal reader: only LIVE links, and
# never an archived (e.g. merged-away) code.
_ensure_temporal_schema()
rows = _db(task_dir).execute(
"""SELECT ac.code_id, ac.started_at, ac.ended_at,
ac.created_by, c.name, c.color, c.parent_id
FROM annotation_codes ac
JOIN codes c ON c.id = ac.code_id
WHERE ac.annotation_id = ?
AND ac.invalidated_at IS NULL
AND c.archived_at IS NULL
ORDER BY c.name ASC""",
(annotation_id,),
).fetchall()
return [dict(r) for r in rows]
# ---- Phase 2 (C): append-only retroactive primitives --------------------
def affected_annotation_ids(
task_dir: str, project: str, code_id: str,
created_by: Optional[str] = None,
) -> List[str]:
"""annotation_ids with a LIVE link to `code_id` (optionally only
those created by `created_by` — the split-by-annotator selector)."""
_ensure_temporal_schema()
q = ("SELECT DISTINCT annotation_id FROM annotation_codes "
"WHERE project = ? AND code_id = ? AND invalidated_at IS NULL")
p: List[Any] = [project, code_id]
if created_by is not None:
q += " AND created_by = ?"
p.append(created_by)
rows = _db(task_dir).execute(q, p).fetchall()
return [r["annotation_id"] for r in rows]
def get_link(
task_dir: str, annotation_id: str, code_id: str
) -> Optional[Dict[str, Any]]:
_ensure_temporal_schema()
row = _db(task_dir).execute(
"""SELECT * FROM annotation_codes
WHERE annotation_id = ? AND code_id = ?""",
(annotation_id, code_id),
).fetchone()
return dict(row) if row else None
def invalidate_links(
task_dir: str, *, project: str, code_id: str, change_id: str,
created_by: Optional[str] = None,
) -> int:
"""Mark live links to `code_id` superseded (append-only — never
DELETE). Optional `created_by` scopes to one annotator (split)."""
_ensure_temporal_schema()
q = ("UPDATE annotation_codes SET invalidated_at = ?, "
"invalidated_by_change = ? "
"WHERE project = ? AND code_id = ? AND invalidated_at IS NULL")
p: List[Any] = [time.time(), change_id, project, code_id]
if created_by is not None:
q += " AND created_by = ?"
p.append(created_by)
conn = _db(task_dir)
cur = conn.execute(q, p)
conn.commit()
return cur.rowcount
def set_link_live(
task_dir: str, *, project: str, annotation_id: str, code_id: str,
created_by: str, started_at: Optional[float] = None,
ended_at: Optional[float] = None,
) -> None:
"""Make (annotation_id, code_id) a LIVE link. Idempotent against the
PK(annotation_id, code_id): if the row exists (live or invalidated)
it is reactivated rather than duplicated/clobbered — this is how a
merge stays correct when the annotation is already on the target."""
_ensure_temporal_schema()
conn = _db(task_dir)
cur = conn.execute(
"""UPDATE annotation_codes
SET invalidated_at = NULL, invalidated_by_change = NULL
WHERE annotation_id = ? AND code_id = ?""",
(annotation_id, code_id),
)
if cur.rowcount == 0:
conn.execute(
"""INSERT INTO annotation_codes
(annotation_id, code_id, project, created_by,
started_at, ended_at)
VALUES (?, ?, ?, ?, ?, ?)""",
(annotation_id, code_id, project, created_by,
started_at, ended_at),
)
conn.commit()
def archive_code(task_dir: str, code_id: str) -> bool:
"""Soft-archive a code (merged away): leaves the live palette + ICL
prompt but the row and its history survive (append-only)."""
_ensure_temporal_schema()
conn = _db(task_dir)
cur = conn.execute(
"UPDATE codes SET archived_at = ?, updated_at = ? WHERE id = ?",
(time.time(), time.time(), code_id),
)
conn.commit()
return cur.rowcount > 0
|