Skip to content

Core API

The core layer provides the fingerprinting engine, similarity scoring, healing logic, and SQLite storage.

ElementFingerprint

Rich identity snapshot of a DOM element.

Attributes:

Name Type Description
tag str

HTML tag name, lowercased (e.g. "button", "input", "div").
text str

Visible text content, stripped and lowercased.
attributes frozenset[tuple[str, str]]

Frozen set of (name, value) pairs for all HTML attributes.
dom_path tuple[str, ...]

Tuple of ancestor tag names from root to this element.
siblings tuple[str, ...]

Tuple of tag names of immediate sibling elements.
bbox BoundingBox | None

Bounding box at fingerprint time, or None if not available.
locator str

The original locator string used to find this element.
test_id str

Identifier for the test that created this fingerprint.
Source code in breadcrumb/core/fingerprint.py 
 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
@dataclass(frozen=True)
class ElementFingerprint:
    """Rich identity snapshot of a DOM element.

    Attributes:
        tag: HTML tag name, lowercased (e.g. "button", "input", "div").
        text: Visible text content, stripped and lowercased.
        attributes: Frozen set of (name, value) pairs for all HTML attributes.
        dom_path: Tuple of ancestor tag names from root to this element.
        siblings: Tuple of tag names of immediate sibling elements.
        bbox: Bounding box at fingerprint time, or None if not available.
        locator: The original locator string used to find this element.
        test_id: Identifier for the test that created this fingerprint.
    """

    tag: str
    text: str
    attributes: frozenset[tuple[str, str]]
    dom_path: tuple[str, ...]
    siblings: tuple[str, ...]
    bbox: BoundingBox | None = None
    locator: str = ""
    test_id: str = ""

    @staticmethod
    def _normalize_text(text: str | None) -> str:
        """Normalize text content: strip, lowercase, collapse whitespace."""
        if not text:
            return ""
        return " ".join(text.strip().lower().split())

    @staticmethod
    def _normalize_tag(tag: str | None) -> str:
        """Normalize tag name: lowercase, strip."""
        if not tag:
            return ""
        return tag.strip().lower()

    @classmethod
    def from_dict(cls, data: dict[str, Any]) -> ElementFingerprint:
        """Create a fingerprint from a plain dictionary.

        Useful for reconstructing fingerprints from storage or test fixtures.
        Handles type coercion (lists -> tuples/frozensets) automatically.

        Args:
            data: Dictionary with fingerprint fields. At minimum requires
                'tag'. All other fields have sensible defaults.

        Returns:
            A new ElementFingerprint instance.
        """
        bbox_data = data.get("bbox")
        bbox: BoundingBox | None = None
        if bbox_data is not None:
            if isinstance(bbox_data, BoundingBox):
                bbox = bbox_data
            elif isinstance(bbox_data, dict):
                bbox = BoundingBox(
                    x=float(bbox_data.get("x", 0)),
                    y=float(bbox_data.get("y", 0)),
                    width=float(bbox_data.get("width", 0)),
                    height=float(bbox_data.get("height", 0)),
                )

        raw_attrs = data.get("attributes", set())
        if isinstance(raw_attrs, dict):
            attrs = frozenset(raw_attrs.items())
        elif isinstance(raw_attrs, (list, tuple)):
            attrs = frozenset(tuple(pair) for pair in raw_attrs)
        elif isinstance(raw_attrs, frozenset):
            attrs = raw_attrs
        else:
            attrs = frozenset()

        raw_dom = data.get("dom_path", ())
        dom_path = tuple(raw_dom) if not isinstance(raw_dom, tuple) else raw_dom

        raw_siblings = data.get("siblings", ())
        siblings = tuple(raw_siblings) if not isinstance(raw_siblings, tuple) else raw_siblings

        return cls(
            tag=cls._normalize_tag(data.get("tag", "")),
            text=cls._normalize_text(data.get("text", "")),
            attributes=attrs,
            dom_path=dom_path,
            siblings=siblings,
            bbox=bbox,
            locator=data.get("locator", ""),
            test_id=data.get("test_id", ""),
        )

    def to_dict(self) -> dict[str, Any]:
        """Serialize the fingerprint to a plain dictionary for storage.

        Returns:
            Dictionary representation suitable for JSON serialization
            and SQLite storage.
        """
        result: dict[str, Any] = {
            "tag": self.tag,
            "text": self.text,
            "attributes": sorted([list(pair) for pair in self.attributes]),
            "dom_path": list(self.dom_path),
            "siblings": list(self.siblings),
            "locator": self.locator,
            "test_id": self.test_id,
        }
        if self.bbox is not None:
            result["bbox"] = {
                "x": self.bbox.x,
                "y": self.bbox.y,
                "width": self.bbox.width,
                "height": self.bbox.height,
            }
        else:
            result["bbox"] = None
        return result

from_dict(data: dict[str, Any]) -> ElementFingerprint

Create a fingerprint from a plain dictionary.

Useful for reconstructing fingerprints from storage or test fixtures. Handles type coercion (lists -> tuples/frozensets) automatically.

Parameters:

Name Type Description Default
data dict[str, Any]

Dictionary with fingerprint fields. At minimum requires 'tag'. All other fields have sensible defaults.

 required

Returns:

Type Description
ElementFingerprint

A new ElementFingerprint instance.
Source code in breadcrumb/core/fingerprint.py 
 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
@classmethod
def from_dict(cls, data: dict[str, Any]) -> ElementFingerprint:
    """Create a fingerprint from a plain dictionary.

    Useful for reconstructing fingerprints from storage or test fixtures.
    Handles type coercion (lists -> tuples/frozensets) automatically.

    Args:
        data: Dictionary with fingerprint fields. At minimum requires
            'tag'. All other fields have sensible defaults.

    Returns:
        A new ElementFingerprint instance.
    """
    bbox_data = data.get("bbox")
    bbox: BoundingBox | None = None
    if bbox_data is not None:
        if isinstance(bbox_data, BoundingBox):
            bbox = bbox_data
        elif isinstance(bbox_data, dict):
            bbox = BoundingBox(
                x=float(bbox_data.get("x", 0)),
                y=float(bbox_data.get("y", 0)),
                width=float(bbox_data.get("width", 0)),
                height=float(bbox_data.get("height", 0)),
            )

    raw_attrs = data.get("attributes", set())
    if isinstance(raw_attrs, dict):
        attrs = frozenset(raw_attrs.items())
    elif isinstance(raw_attrs, (list, tuple)):
        attrs = frozenset(tuple(pair) for pair in raw_attrs)
    elif isinstance(raw_attrs, frozenset):
        attrs = raw_attrs
    else:
        attrs = frozenset()

    raw_dom = data.get("dom_path", ())
    dom_path = tuple(raw_dom) if not isinstance(raw_dom, tuple) else raw_dom

    raw_siblings = data.get("siblings", ())
    siblings = tuple(raw_siblings) if not isinstance(raw_siblings, tuple) else raw_siblings

    return cls(
        tag=cls._normalize_tag(data.get("tag", "")),
        text=cls._normalize_text(data.get("text", "")),
        attributes=attrs,
        dom_path=dom_path,
        siblings=siblings,
        bbox=bbox,
        locator=data.get("locator", ""),
        test_id=data.get("test_id", ""),
    )

to_dict() -> dict[str, Any]

Serialize the fingerprint to a plain dictionary for storage.

Returns:

Type Description
dict[str, Any]

Dictionary representation suitable for JSON serialization
dict[str, Any]

and SQLite storage.
Source code in breadcrumb/core/fingerprint.py 
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
def to_dict(self) -> dict[str, Any]:
    """Serialize the fingerprint to a plain dictionary for storage.

    Returns:
        Dictionary representation suitable for JSON serialization
        and SQLite storage.
    """
    result: dict[str, Any] = {
        "tag": self.tag,
        "text": self.text,
        "attributes": sorted([list(pair) for pair in self.attributes]),
        "dom_path": list(self.dom_path),
        "siblings": list(self.siblings),
        "locator": self.locator,
        "test_id": self.test_id,
    }
    if self.bbox is not None:
        result["bbox"] = {
            "x": self.bbox.x,
            "y": self.bbox.y,
            "width": self.bbox.width,
            "height": self.bbox.height,
        }
    else:
        result["bbox"] = None
    return result

BoundingBox

Axis-aligned bounding box of an element's visual position.

Source code in breadcrumb/core/fingerprint.py 
22
23
24
25
26
27
28
29
30
31
32
33
34
@dataclass(frozen=True)
class BoundingBox:
    """Axis-aligned bounding box of an element's visual position."""

    x: float
    y: float
    width: float
    height: float

    @property
    def center(self) -> tuple[float, float]:
        """Return the center point of the bounding box."""
        return (self.x + self.width / 2, self.y + self.height / 2)

center: tuple[float, float]

Return the center point of the bounding box.

FingerprintStore

SQLite-backed storage for fingerprints and healing events.

Usage::

store = FingerprintStore()          # Uses .breadcrumb.db in cwd
store = FingerprintStore("path/to/db.sqlite")
store.save_fingerprint(fingerprint)
fp = store.load_fingerprint("test_login", "#login-btn")

The database is created automatically on first access.

Source code in breadcrumb/core/storage.py 
 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
class FingerprintStore:
    """SQLite-backed storage for fingerprints and healing events.

    Usage::

        store = FingerprintStore()          # Uses .breadcrumb.db in cwd
        store = FingerprintStore("path/to/db.sqlite")
        store.save_fingerprint(fingerprint)
        fp = store.load_fingerprint("test_login", "#login-btn")

    The database is created automatically on first access.
    """

    def __init__(self, db_path: str | Path = DEFAULT_DB_PATH) -> None:
        self._db_path = Path(db_path)
        self._conn: sqlite3.Connection | None = None
        self._ensure_db()

    @property
    def db_path(self) -> Path:
        """Return the path to the database file."""
        return self._db_path

    def _get_conn(self) -> sqlite3.Connection:
        """Get or create the database connection."""
        if self._conn is None:
            self._conn = sqlite3.connect(str(self._db_path))
            self._conn.row_factory = sqlite3.Row
            self._conn.execute("PRAGMA journal_mode=WAL")
            self._conn.execute("PRAGMA foreign_keys=ON")
        return self._conn

    def _ensure_db(self) -> None:
        """Create tables if they don't exist."""
        conn = self._get_conn()
        conn.executescript(_CREATE_TABLES_SQL)

        row = conn.execute(
            "SELECT value FROM schema_meta WHERE key = 'schema_version'",
        ).fetchone()
        if row is None:
            conn.execute(
                "INSERT INTO schema_meta (key, value) VALUES ('schema_version', ?)",
                (str(SCHEMA_VERSION),),
            )
            conn.commit()

    def save_fingerprint(self, fingerprint: ElementFingerprint) -> None:
        """Save or update a fingerprint for a (test_id, locator) pair.

        Called on passing test runs to keep the fingerprint database current.

        Raises:
            ValueError: If test_id or locator is empty.
        """
        if not fingerprint.test_id:
            msg = "Fingerprint must have a test_id to be stored."
            raise ValueError(msg)
        if not fingerprint.locator:
            msg = "Fingerprint must have a locator to be stored."
            raise ValueError(msg)

        conn = self._get_conn()
        conn.execute(
            "INSERT INTO fingerprints (test_id, locator, fingerprint_json, updated_at) "
            "VALUES (?, ?, ?, ?) "
            "ON CONFLICT (test_id, locator) "
            "DO UPDATE SET fingerprint_json = excluded.fingerprint_json, "
            "             updated_at = excluded.updated_at",
            (
                fingerprint.test_id,
                fingerprint.locator,
                json.dumps(fingerprint.to_dict()),
                time.time(),
            ),
        )
        conn.commit()

    def load_fingerprint(self, test_id: str, locator: str) -> ElementFingerprint | None:
        """Load a stored fingerprint for a (test_id, locator) pair.

        Returns None if not found.
        """
        conn = self._get_conn()
        row = conn.execute(
            "SELECT fingerprint_json FROM fingerprints WHERE test_id = ? AND locator = ?",
            (test_id, locator),
        ).fetchone()

        if row is None:
            return None

        data: dict[str, Any] = json.loads(row[0])  # type: ignore[index]
        return ElementFingerprint.from_dict(data)

    def record_healing(self, event: HealingEvent) -> None:
        """Record a healing event. Append-only for reporting."""
        conn = self._get_conn()
        conn.execute(
            "INSERT INTO healing_events "
            "(test_id, locator, confidence, original_json, healed_json, timestamp) "
            "VALUES (?, ?, ?, ?, ?, ?)",
            (
                event.test_id,
                event.locator,
                event.confidence,
                json.dumps(event.original_fingerprint),
                json.dumps(event.healed_fingerprint),
                event.timestamp,
            ),
        )
        conn.commit()

    def get_healing_events(
        self,
        test_id: str | None = None,
        locator: str | None = None,
    ) -> list[HealingEvent]:
        """Query healing events, optionally filtered by test and/or locator.

        Returns list ordered by timestamp descending.
        """
        conn = self._get_conn()
        query = "SELECT * FROM healing_events"
        params: list[str] = []
        conditions: list[str] = []

        if test_id is not None:
            conditions.append("test_id = ?")
            params.append(test_id)
        if locator is not None:
            conditions.append("locator = ?")
            params.append(locator)

        if conditions:
            query += " WHERE " + " AND ".join(conditions)

        query += " ORDER BY timestamp DESC"

        rows = conn.execute(query, params).fetchall()
        return [
            HealingEvent(
                test_id=r["test_id"],  # type: ignore[index]
                locator=r["locator"],  # type: ignore[index]
                confidence=r["confidence"],  # type: ignore[index]
                original_fingerprint=json.loads(r["original_json"]),  # type: ignore[index]
                healed_fingerprint=json.loads(r["healed_json"]),  # type: ignore[index]
                timestamp=r["timestamp"],  # type: ignore[index]
            )
            for r in rows
        ]

    def get_all_fingerprints(self) -> list[ElementFingerprint]:
        """Load all stored fingerprints."""
        conn = self._get_conn()
        rows = conn.execute("SELECT fingerprint_json FROM fingerprints").fetchall()
        return [
            ElementFingerprint.from_dict(json.loads(r[0]))  # type: ignore[index]
            for r in rows
        ]

    def delete_fingerprint(self, test_id: str, locator: str) -> bool:
        """Delete a stored fingerprint. Returns True if found and deleted."""
        conn = self._get_conn()
        cursor = conn.execute(
            "DELETE FROM fingerprints WHERE test_id = ? AND locator = ?",
            (test_id, locator),
        )
        conn.commit()
        return cursor.rowcount > 0

    def clear(self) -> None:
        """Delete all fingerprints and healing events."""
        conn = self._get_conn()
        conn.execute("DELETE FROM fingerprints")
        conn.execute("DELETE FROM healing_events")
        conn.commit()

    def stats(self) -> dict[str, int]:
        """Return counts of fingerprints and healing events."""
        conn = self._get_conn()
        fp_row = conn.execute("SELECT COUNT(*) FROM fingerprints").fetchone()
        he_row = conn.execute("SELECT COUNT(*) FROM healing_events").fetchone()
        fp_count: int = fp_row[0] if fp_row is not None else 0  # type: ignore[index]
        he_count: int = he_row[0] if he_row is not None else 0  # type: ignore[index]
        return {"fingerprints": fp_count, "healing_events": he_count}

    def close(self) -> None:
        """Close the database connection."""
        if self._conn is not None:
            self._conn.close()
            self._conn = None

db_path: Path

Return the path to the database file.

save_fingerprint(fingerprint: ElementFingerprint) -> None

Save or update a fingerprint for a (test_id, locator) pair.

Called on passing test runs to keep the fingerprint database current.

Raises:

Type Description
ValueError

If test_id or locator is empty.
Source code in breadcrumb/core/storage.py 
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
def save_fingerprint(self, fingerprint: ElementFingerprint) -> None:
    """Save or update a fingerprint for a (test_id, locator) pair.

    Called on passing test runs to keep the fingerprint database current.

    Raises:
        ValueError: If test_id or locator is empty.
    """
    if not fingerprint.test_id:
        msg = "Fingerprint must have a test_id to be stored."
        raise ValueError(msg)
    if not fingerprint.locator:
        msg = "Fingerprint must have a locator to be stored."
        raise ValueError(msg)

    conn = self._get_conn()
    conn.execute(
        "INSERT INTO fingerprints (test_id, locator, fingerprint_json, updated_at) "
        "VALUES (?, ?, ?, ?) "
        "ON CONFLICT (test_id, locator) "
        "DO UPDATE SET fingerprint_json = excluded.fingerprint_json, "
        "             updated_at = excluded.updated_at",
        (
            fingerprint.test_id,
            fingerprint.locator,
            json.dumps(fingerprint.to_dict()),
            time.time(),
        ),
    )
    conn.commit()

load_fingerprint(test_id: str, locator: str) -> ElementFingerprint | None

Load a stored fingerprint for a (test_id, locator) pair.

Returns None if not found.

Source code in breadcrumb/core/storage.py 
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
def load_fingerprint(self, test_id: str, locator: str) -> ElementFingerprint | None:
    """Load a stored fingerprint for a (test_id, locator) pair.

    Returns None if not found.
    """
    conn = self._get_conn()
    row = conn.execute(
        "SELECT fingerprint_json FROM fingerprints WHERE test_id = ? AND locator = ?",
        (test_id, locator),
    ).fetchone()

    if row is None:
        return None

    data: dict[str, Any] = json.loads(row[0])  # type: ignore[index]
    return ElementFingerprint.from_dict(data)

record_healing(event: HealingEvent) -> None

Record a healing event. Append-only for reporting.

Source code in breadcrumb/core/storage.py 
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
def record_healing(self, event: HealingEvent) -> None:
    """Record a healing event. Append-only for reporting."""
    conn = self._get_conn()
    conn.execute(
        "INSERT INTO healing_events "
        "(test_id, locator, confidence, original_json, healed_json, timestamp) "
        "VALUES (?, ?, ?, ?, ?, ?)",
        (
            event.test_id,
            event.locator,
            event.confidence,
            json.dumps(event.original_fingerprint),
            json.dumps(event.healed_fingerprint),
            event.timestamp,
        ),
    )
    conn.commit()

get_healing_events(test_id: str | None = None, locator: str | None = None) -> list[HealingEvent]

Query healing events, optionally filtered by test and/or locator.

Returns list ordered by timestamp descending.

Source code in breadcrumb/core/storage.py 
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
def get_healing_events(
    self,
    test_id: str | None = None,
    locator: str | None = None,
) -> list[HealingEvent]:
    """Query healing events, optionally filtered by test and/or locator.

    Returns list ordered by timestamp descending.
    """
    conn = self._get_conn()
    query = "SELECT * FROM healing_events"
    params: list[str] = []
    conditions: list[str] = []

    if test_id is not None:
        conditions.append("test_id = ?")
        params.append(test_id)
    if locator is not None:
        conditions.append("locator = ?")
        params.append(locator)

    if conditions:
        query += " WHERE " + " AND ".join(conditions)

    query += " ORDER BY timestamp DESC"

    rows = conn.execute(query, params).fetchall()
    return [
        HealingEvent(
            test_id=r["test_id"],  # type: ignore[index]
            locator=r["locator"],  # type: ignore[index]
            confidence=r["confidence"],  # type: ignore[index]
            original_fingerprint=json.loads(r["original_json"]),  # type: ignore[index]
            healed_fingerprint=json.loads(r["healed_json"]),  # type: ignore[index]
            timestamp=r["timestamp"],  # type: ignore[index]
        )
        for r in rows
    ]

get_all_fingerprints() -> list[ElementFingerprint]

Load all stored fingerprints.

Source code in breadcrumb/core/storage.py 
222
223
224
225
226
227
228
229
def get_all_fingerprints(self) -> list[ElementFingerprint]:
    """Load all stored fingerprints."""
    conn = self._get_conn()
    rows = conn.execute("SELECT fingerprint_json FROM fingerprints").fetchall()
    return [
        ElementFingerprint.from_dict(json.loads(r[0]))  # type: ignore[index]
        for r in rows
    ]

delete_fingerprint(test_id: str, locator: str) -> bool

Delete a stored fingerprint. Returns True if found and deleted.

Source code in breadcrumb/core/storage.py 
231
232
233
234
235
236
237
238
239
def delete_fingerprint(self, test_id: str, locator: str) -> bool:
    """Delete a stored fingerprint. Returns True if found and deleted."""
    conn = self._get_conn()
    cursor = conn.execute(
        "DELETE FROM fingerprints WHERE test_id = ? AND locator = ?",
        (test_id, locator),
    )
    conn.commit()
    return cursor.rowcount > 0

clear() -> None

Delete all fingerprints and healing events.

Source code in breadcrumb/core/storage.py 
241
242
243
244
245
246
def clear(self) -> None:
    """Delete all fingerprints and healing events."""
    conn = self._get_conn()
    conn.execute("DELETE FROM fingerprints")
    conn.execute("DELETE FROM healing_events")
    conn.commit()

stats() -> dict[str, int]

Return counts of fingerprints and healing events.

Source code in breadcrumb/core/storage.py 
248
249
250
251
252
253
254
255
def stats(self) -> dict[str, int]:
    """Return counts of fingerprints and healing events."""
    conn = self._get_conn()
    fp_row = conn.execute("SELECT COUNT(*) FROM fingerprints").fetchone()
    he_row = conn.execute("SELECT COUNT(*) FROM healing_events").fetchone()
    fp_count: int = fp_row[0] if fp_row is not None else 0  # type: ignore[index]
    he_count: int = he_row[0] if he_row is not None else 0  # type: ignore[index]
    return {"fingerprints": fp_count, "healing_events": he_count}

close() -> None

Close the database connection.

Source code in breadcrumb/core/storage.py 
257
258
259
260
261
def close(self) -> None:
    """Close the database connection."""
    if self._conn is not None:
        self._conn.close()
        self._conn = None

HealingEvent

Record of a single healing occurrence.

Source code in breadcrumb/core/storage.py 
58
59
60
61
62
63
64
65
66
67
@dataclass
class HealingEvent:
    """Record of a single healing occurrence."""

    test_id: str
    locator: str
    confidence: float
    original_fingerprint: dict[str, Any]
    healed_fingerprint: dict[str, Any]
    timestamp: float