Skip to content

profile

Portable timing profile storage for cross-run/machine estimation.

Classes

RuleProfile dataclass

Timing profile for a single rule.

Attributes:

Name Type Description
rule str

The rule name.
sample_count int

Number of executions observed.
mean_duration float

Mean duration in seconds.
std_dev float

Standard deviation of durations.
min_duration float

Minimum observed duration.
max_duration float

Maximum observed duration.
durations list[float]

Raw duration values for merging.
timestamps list[float]

Unix timestamps when each run completed.
Source code in snakesee/profile.py 
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
@dataclass
class RuleProfile:
    """
    Timing profile for a single rule.

    Attributes:
        rule: The rule name.
        sample_count: Number of executions observed.
        mean_duration: Mean duration in seconds.
        std_dev: Standard deviation of durations.
        min_duration: Minimum observed duration.
        max_duration: Maximum observed duration.
        durations: Raw duration values for merging.
        timestamps: Unix timestamps when each run completed.
    """

    rule: str
    sample_count: int
    mean_duration: float
    std_dev: float
    min_duration: float
    max_duration: float
    durations: list[float] = field(default_factory=list)
    timestamps: list[float] = field(default_factory=list)

    @classmethod
    def from_stats(cls, stats: RuleTimingStats) -> RuleProfile:
        """Create a RuleProfile from RuleTimingStats."""
        return cls(
            rule=stats.rule,
            sample_count=stats.count,
            mean_duration=stats.mean_duration,
            std_dev=stats.std_dev,
            min_duration=stats.min_duration,
            max_duration=stats.max_duration,
            durations=list(stats.durations),
            timestamps=list(stats.timestamps),
        )

    def to_stats(self) -> RuleTimingStats:
        """Convert back to RuleTimingStats."""
        return RuleTimingStats(
            rule=self.rule,
            durations=list(self.durations),
            timestamps=list(self.timestamps),
        )

Functions

from_stats classmethod 
from_stats(stats: RuleTimingStats) -> RuleProfile

Create a RuleProfile from RuleTimingStats.

Source code in snakesee/profile.py 
47
48
49
50
51
52
53
54
55
56
57
58
59
@classmethod
def from_stats(cls, stats: RuleTimingStats) -> RuleProfile:
    """Create a RuleProfile from RuleTimingStats."""
    return cls(
        rule=stats.rule,
        sample_count=stats.count,
        mean_duration=stats.mean_duration,
        std_dev=stats.std_dev,
        min_duration=stats.min_duration,
        max_duration=stats.max_duration,
        durations=list(stats.durations),
        timestamps=list(stats.timestamps),
    )
to_stats 
to_stats() -> RuleTimingStats

Convert back to RuleTimingStats.

Source code in snakesee/profile.py 
61
62
63
64
65
66
67
def to_stats(self) -> RuleTimingStats:
    """Convert back to RuleTimingStats."""
    return RuleTimingStats(
        rule=self.rule,
        durations=list(self.durations),
        timestamps=list(self.timestamps),
    )

TimingProfile dataclass

Complete timing profile for a workflow.

Attributes:

Name Type Description
version int

Profile format version.
created str

ISO timestamp when profile was first created.
updated str

ISO timestamp when profile was last updated.
machine str | None

Optional machine identifier.
rules dict[str, RuleProfile]

Dictionary of rule profiles.
Source code in snakesee/profile.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
@dataclass
class TimingProfile:
    """
    Complete timing profile for a workflow.

    Attributes:
        version: Profile format version.
        created: ISO timestamp when profile was first created.
        updated: ISO timestamp when profile was last updated.
        machine: Optional machine identifier.
        rules: Dictionary of rule profiles.
    """

    version: int
    created: str
    updated: str
    rules: dict[str, RuleProfile]
    machine: str | None = None

    @classmethod
    def create_new(cls, rule_stats: dict[str, RuleTimingStats]) -> TimingProfile:
        """Create a new profile from rule timing stats."""
        now = datetime.now(timezone.utc).isoformat()
        rules = {name: RuleProfile.from_stats(stats) for name, stats in rule_stats.items()}
        return cls(
            version=PROFILE_VERSION,
            created=now,
            updated=now,
            machine=socket.gethostname(),
            rules=rules,
        )

    def to_rule_stats(self) -> dict[str, RuleTimingStats]:
        """Convert profile back to rule timing stats."""
        return {name: profile.to_stats() for name, profile in self.rules.items()}

    def merge_with(self, other: TimingProfile) -> TimingProfile:
        """
        Merge another profile into this one.

        Combines durations and timestamps from both profiles, keeping all data
        for weighted estimation. Updates the 'updated' timestamp.

        Args:
            other: Another profile to merge.

        Returns:
            A new merged profile.
        """
        merged_rules: dict[str, RuleProfile] = {}

        # Start with rules from self
        all_rule_names = set(self.rules.keys()) | set(other.rules.keys())

        for rule_name in all_rule_names:
            self_profile = self.rules.get(rule_name)
            other_profile = other.rules.get(rule_name)

            if self_profile is None and other_profile is not None:
                merged_rules[rule_name] = other_profile
            elif other_profile is None and self_profile is not None:
                merged_rules[rule_name] = self_profile
            elif self_profile is not None and other_profile is not None:
                # Merge durations and timestamps, then sort by timestamp
                combined = list(zip(self_profile.durations, self_profile.timestamps, strict=True))
                combined.extend(zip(other_profile.durations, other_profile.timestamps, strict=True))
                combined.sort(key=lambda x: x[1])  # Sort by timestamp

                durations = [x[0] for x in combined]
                timestamps = [x[1] for x in combined]

                # Create merged stats to get computed properties
                merged_stats = RuleTimingStats(
                    rule=rule_name,
                    durations=durations,
                    timestamps=timestamps,
                )

                merged_rules[rule_name] = RuleProfile(
                    rule=rule_name,
                    sample_count=merged_stats.count,
                    mean_duration=merged_stats.mean_duration,
                    std_dev=merged_stats.std_dev,
                    min_duration=merged_stats.min_duration,
                    max_duration=merged_stats.max_duration,
                    durations=durations,
                    timestamps=timestamps,
                )

        now = datetime.now(timezone.utc).isoformat()
        return TimingProfile(
            version=PROFILE_VERSION,
            created=self.created,  # Keep original creation time
            updated=now,
            machine=self.machine,  # Keep original machine
            rules=merged_rules,
        )

Functions

create_new classmethod 
create_new(rule_stats: dict[str, RuleTimingStats]) -> TimingProfile

Create a new profile from rule timing stats.

Source code in snakesee/profile.py 
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
@classmethod
def create_new(cls, rule_stats: dict[str, RuleTimingStats]) -> TimingProfile:
    """Create a new profile from rule timing stats."""
    now = datetime.now(timezone.utc).isoformat()
    rules = {name: RuleProfile.from_stats(stats) for name, stats in rule_stats.items()}
    return cls(
        version=PROFILE_VERSION,
        created=now,
        updated=now,
        machine=socket.gethostname(),
        rules=rules,
    )
merge_with 
merge_with(other: TimingProfile) -> TimingProfile

Merge another profile into this one.

Combines durations and timestamps from both profiles, keeping all data for weighted estimation. Updates the 'updated' timestamp.

Parameters:

Name Type Description Default
other TimingProfile

Another profile to merge.

 required

Returns:

Type Description
TimingProfile

A new merged profile.
Source code in snakesee/profile.py 
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
def merge_with(self, other: TimingProfile) -> TimingProfile:
    """
    Merge another profile into this one.

    Combines durations and timestamps from both profiles, keeping all data
    for weighted estimation. Updates the 'updated' timestamp.

    Args:
        other: Another profile to merge.

    Returns:
        A new merged profile.
    """
    merged_rules: dict[str, RuleProfile] = {}

    # Start with rules from self
    all_rule_names = set(self.rules.keys()) | set(other.rules.keys())

    for rule_name in all_rule_names:
        self_profile = self.rules.get(rule_name)
        other_profile = other.rules.get(rule_name)

        if self_profile is None and other_profile is not None:
            merged_rules[rule_name] = other_profile
        elif other_profile is None and self_profile is not None:
            merged_rules[rule_name] = self_profile
        elif self_profile is not None and other_profile is not None:
            # Merge durations and timestamps, then sort by timestamp
            combined = list(zip(self_profile.durations, self_profile.timestamps, strict=True))
            combined.extend(zip(other_profile.durations, other_profile.timestamps, strict=True))
            combined.sort(key=lambda x: x[1])  # Sort by timestamp

            durations = [x[0] for x in combined]
            timestamps = [x[1] for x in combined]

            # Create merged stats to get computed properties
            merged_stats = RuleTimingStats(
                rule=rule_name,
                durations=durations,
                timestamps=timestamps,
            )

            merged_rules[rule_name] = RuleProfile(
                rule=rule_name,
                sample_count=merged_stats.count,
                mean_duration=merged_stats.mean_duration,
                std_dev=merged_stats.std_dev,
                min_duration=merged_stats.min_duration,
                max_duration=merged_stats.max_duration,
                durations=durations,
                timestamps=timestamps,
            )

    now = datetime.now(timezone.utc).isoformat()
    return TimingProfile(
        version=PROFILE_VERSION,
        created=self.created,  # Keep original creation time
        updated=now,
        machine=self.machine,  # Keep original machine
        rules=merged_rules,
    )
to_rule_stats 
to_rule_stats() -> dict[str, RuleTimingStats]

Convert profile back to rule timing stats.

Source code in snakesee/profile.py 
102
103
104
def to_rule_stats(self) -> dict[str, RuleTimingStats]:
    """Convert profile back to rule timing stats."""
    return {name: profile.to_stats() for name, profile in self.rules.items()}

Functions

export_profile_from_metadata 

export_profile_from_metadata(metadata_dir: Path, output_path: Path | None = None, merge_existing: bool = False) -> TimingProfile

Export a timing profile from Snakemake metadata.

Parameters:

Name Type Description Default
metadata_dir Path

Path to .snakemake/metadata/ directory.

 required
output_path Path | None

Path to write the profile. If None, returns without saving.

 None
merge_existing bool

If True and output_path exists, merge with existing profile.

 False

Returns:

Type Description
TimingProfile

The created/merged profile.
Source code in snakesee/profile.py 
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
def export_profile_from_metadata(
    metadata_dir: Path,
    output_path: Path | None = None,
    merge_existing: bool = False,
) -> TimingProfile:
    """
    Export a timing profile from Snakemake metadata.

    Args:
        metadata_dir: Path to .snakemake/metadata/ directory.
        output_path: Path to write the profile. If None, returns without saving.
        merge_existing: If True and output_path exists, merge with existing profile.

    Returns:
        The created/merged profile.
    """
    from snakesee.parser import collect_rule_timing_stats

    rule_stats = collect_rule_timing_stats(metadata_dir)
    new_profile = TimingProfile.create_new(rule_stats)

    if merge_existing and output_path is not None and output_path.exists():
        existing = load_profile(output_path)
        new_profile = existing.merge_with(new_profile)

    if output_path is not None:
        save_profile(new_profile, output_path)

    return new_profile

find_profile 

find_profile(workflow_dir: Path) -> Path | None

Search for a profile file in the workflow directory and parent directories.

Searches for .snakesee-profile.json in: 1. workflow_dir 2. Parent directories (up to 5 levels)

Parameters:

Name Type Description Default
workflow_dir Path

Starting directory to search from.

 required

Returns:

Type Description
Path | None

Path to the found profile, or None if not found.
Source code in snakesee/profile.py 
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
def find_profile(workflow_dir: Path) -> Path | None:
    """
    Search for a profile file in the workflow directory and parent directories.

    Searches for .snakesee-profile.json in:
    1. workflow_dir
    2. Parent directories (up to 5 levels)

    Args:
        workflow_dir: Starting directory to search from.

    Returns:
        Path to the found profile, or None if not found.
    """
    current = workflow_dir.resolve()
    for _ in range(6):  # Current + 5 parent levels
        profile_path = current / DEFAULT_PROFILE_NAME
        if profile_path.exists():
            return profile_path
        if current.parent == current:
            break
        current = current.parent
    return None

load_profile 

load_profile(path: Path) -> TimingProfile

Load a timing profile from a JSON file.

Parameters:

Name Type Description Default
path Path

Path to the profile file.

 required

Returns:

Type Description
TimingProfile

The loaded profile.

Raises:

Type Description
ProfileNotFoundError

If the profile doesn't exist.
InvalidProfileError

If the profile format is invalid.
Source code in snakesee/profile.py 
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
def load_profile(path: Path) -> TimingProfile:
    """
    Load a timing profile from a JSON file.

    Args:
        path: Path to the profile file.

    Returns:
        The loaded profile.

    Raises:
        ProfileNotFoundError: If the profile doesn't exist.
        InvalidProfileError: If the profile format is invalid.
    """
    from snakesee.exceptions import InvalidProfileError
    from snakesee.exceptions import ProfileNotFoundError

    try:
        data = json.loads(path.read_text())
    except FileNotFoundError as e:
        raise ProfileNotFoundError(path) from e
    except json.JSONDecodeError as e:
        raise InvalidProfileError(path, "Invalid JSON format", cause=e) from e

    version = data.get("version", 1)
    if version > PROFILE_VERSION:
        raise InvalidProfileError(
            path,
            f"Profile version {version} is newer than supported version {PROFILE_VERSION}",
        )

    try:
        rules = {}
        for name, rule_data in data.get("rules", {}).items():
            rules[name] = RuleProfile(
                rule=rule_data["rule"],
                sample_count=rule_data["sample_count"],
                mean_duration=rule_data["mean_duration"],
                std_dev=rule_data["std_dev"],
                min_duration=rule_data["min_duration"],
                max_duration=rule_data["max_duration"],
                durations=rule_data.get("durations", []),
                timestamps=rule_data.get("timestamps", []),
            )

        return TimingProfile(
            version=version,
            created=data["created"],
            updated=data["updated"],
            machine=data.get("machine"),
            rules=rules,
        )
    except KeyError as e:
        raise InvalidProfileError(path, f"Missing required field: {e}") from e
    except TypeError as e:
        raise InvalidProfileError(path, f"Invalid data structure: {e}") from e

save_profile 

save_profile(profile: TimingProfile, path: Path) -> None

Save a timing profile to a JSON file.

Parameters:

Name Type Description Default
profile TimingProfile

The profile to save.

 required
path Path

Path to write the profile to.

 required
Source code in snakesee/profile.py 
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
def save_profile(profile: TimingProfile, path: Path) -> None:
    """
    Save a timing profile to a JSON file.

    Args:
        profile: The profile to save.
        path: Path to write the profile to.
    """
    data = {
        "version": profile.version,
        "created": profile.created,
        "updated": profile.updated,
        "machine": profile.machine,
        "rules": {name: asdict(rp) for name, rp in profile.rules.items()},
    }

    # Write atomically: write to temp file, then rename
    # This prevents corruption if the program crashes mid-write
    content = json.dumps(data, indent=2)
    fd, temp_path = tempfile.mkstemp(
        suffix=".tmp",
        prefix=path.name,
        dir=path.parent,
    )
    try:
        with os.fdopen(fd, "w", encoding="utf-8") as f:
            f.write(content)
        os.replace(temp_path, path)
    except Exception:
        # Clean up temp file on any error
        try:
            os.unlink(temp_path)
        except OSError:
            pass
        raise