mirror of
https://github.com/priyanshujain/sanderling.git
synced 2026-10-02 11:07:10 +00:00
* feat(cli): add --max-steps for step-bounded runs runner.Options.MaxSteps already worked but was unreachable from the command line. A step budget is what makes two generators comparable: one making a model call per step and one drawing from a PRNG are not comparable per second. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * feat(trace): record arm membership and host in meta.json meta.json recorded the seed but not which picker ran, how it was configured, what budget it was given, or which machine produced it. A directory of runs cannot be attributed to an experiment cell without those, which makes any factorial computed from such a directory unanalysable after the fact. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * feat(cli): add --arm and populate run meta from it Model and instructions are recorded only when the LLM picker is the one that will actually run, so a spec declaring generator = llm() that is run under the seeded picker does not label its trace with a model it never called. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * feat(campaign): sweep seeds for one experiment cell campaign.json lists the seeds a sweep intended to run and is written before the first run, so a host that dropped runs shows up as missing seeds rather than as a smaller sample. Seed 0 is rejected: sanderling test reads it as "derive a seed from the clock", which is why conformance/gates.sh controls nothing today. Each run contributes one runs.jsonl line carrying steps to first violation by origin step, the step that armed the failed obligation, so the survival analysis never reopens a trace. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * fix(runner): no silent generator fallback, and llm on web --generator llm against a spec declaring no generator = llm(...) logged a warning and ran the seeded picker. For a comparison campaign that is silent arm corruption: the run completes, the directory looks correct, and the wrong policy drove it. It is now fatal. pickSources also returned the V8 source for both action and extractor on web before it looked at the generator, so the llm policy was unreachable there. The two axes are now independent: the driver picks the extractor source, the flag picks the action source, and llmSource composes with either because the runner populates the candidate list and screenshot on every platform. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * fix(chrome): make the hierarchy dump agree with the web runtime Three facts differed between the dump the goja host reads and the DOM the V8 host reads, so the two enumerated different candidates on one page. scrollable was never emitted, and worker.go reads exactly that attribute while targets.ts requires it for scrolls, so the goja host could not offer a single web scroll. clickable tested el.onclick, which React assigns to its root container for event delegation, making the whole viewport a tap target here and in no other enumeration. Both now resolve through the selector sets in pkg/spec/src/web-runtime.ts. The dump also rooted at body while collectTargets walks querySelectorAll("*"), so the goja host never saw html, where page-level scrolling lives. It now roots at documentElement and skips the head subtree, which is all zero-bounds and would otherwise carry script and title text into the trace. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * fix(conformance): give the gate reproducible seeds SEED defaulted to 0 and sanderling test reads --seed 0 as "derive a seed from the clock", so the tunable controlled nothing and a gate failure could not be re-run. SEEDS now takes one explicit non-zero seed per run, recorded in the results table so a failing row names its stream. The five runs stay on five different streams: a gate that scored one path five times would catch less than one that scores five. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * fix(chrome): emit editable as a plain boolean editable was emitted as `isEditable || null`, and an absent field sends internal/hierarchy into the native fallback, which reads any class name containing "EditText" as an Android text widget. On web that is just a CSS class, so a page styling a div with it was editable to the goja host and not to the web runtime, and the model policy could be offered typing into a div. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * fix(spec): leave the head subtree out of the web target walk collectTargets walked querySelectorAll("*") while the hierarchy dump skips head, so the two hosts enumerated different element sets on every page with a <head>. No candidate changes: builtinCandidates pushes only for targets acceptsTarget admits, and head elements have no positive bounds, so the list the draw ranges over is untouched. What changes is that targetIndex now means the same thing on both hosts. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * test(chrome): compare the facts both hosts derive from one DOM The existing parity harness hand-authors the facts on both sides, so it proves that given identical facts both hosts select identical candidates, and says nothing about the two code paths that derive those facts from a real page. Four divergences lived in that blind spot and it passed throughout. This drives one real page and compares clickable, enabled, editable, scrollable and positiveBounds element by element, plus the element sets themselves, which is what catches a host that omits html or includes head. Reverting any of the four fixes makes it fail naming the element and the fact. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * chore(make): run the browser packages one at a time Both launch Chrome and launching two at once has failed with "Launch: context canceled". Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * style: remove every em-dash and en-dash Eighteen occurrences across fourteen files. Each sentence was repunctuated to suit what the dash was doing rather than swapped for a hyphen, which produces comma splices. The minus sign in folio-web's ledger is a minus sign and stays. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * fix(chrome): honor the caller context in Launch Launch and clearState ran against d.tabCtx, so a target that accepts the connection and never answers wedged the process past its own --duration and through SIGTERM, needing SIGKILL. Unattended that is a campaign worker lost for the rest of the sweep with no diagnostic. The browser is still allocated against d.tabCtx first, because chromedp starts Chrome under whichever context calls Run first and allocating under a caller deadline would kill the browser when Launch returns. Everything after allocation goes through runCtx. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * fix(sidecarassets): publish the extracted jar through a rename Extract wrote a 96 MB jar with a plain WriteFile into a temp path every sanderling process on the host shares. On a cold host several concurrent workers all miss the checksum and all write the same path, and O_TRUNC lets one spawn a JVM against another's half-written archive. A fresh experiment host is exactly a cold host. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * feat(campaign): kill a run that outlives --run-timeout A wedged run holds its worker for the rest of the sweep, and on an unattended host nothing else will send it a signal. Defaults to three times --duration and must exceed it. A killed run is recorded as timed_out rather than as a generic failure, so the analysis can tell a lost cell from a real crash. Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX * style(test): gofmt browser_test.go Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX
221 lines
8.4 KiB
Go
221 lines
8.4 KiB
Go
// Package trace records each run's steps, snapshots, and violations to disk for later inspection.
|
|
package trace
|
|
|
|
import (
|
|
"encoding/json"
|
|
"fmt"
|
|
"io"
|
|
"os"
|
|
"path/filepath"
|
|
"sync"
|
|
"time"
|
|
|
|
"github.com/priyanshujain/sanderling/internal/hierarchy"
|
|
)
|
|
|
|
type Step struct {
|
|
Index int `json:"step"`
|
|
Timestamp time.Time `json:"timestamp"`
|
|
Screen string `json:"screen,omitempty"`
|
|
Snapshots map[string]json.RawMessage `json:"snapshots,omitempty"`
|
|
// NextAction is the action chosen for the next iteration based on observing this step.
|
|
NextAction *Action `json:"next_action,omitempty"`
|
|
Exceptions []Exception `json:"exceptions,omitempty"`
|
|
Violations []string `json:"violations,omitempty"`
|
|
Hierarchy *hierarchy.Tree `json:"hierarchy,omitempty"`
|
|
Residuals map[string]json.RawMessage `json:"residuals,omitempty"`
|
|
Metrics *Metrics `json:"metrics,omitempty"`
|
|
ExtractorChanges map[string]ExtractorChange `json:"extractor_changes,omitempty"`
|
|
// Transitional marks a step whose hierarchy still showed a NavHost
|
|
// cross-fade (multiple route-level *Screen ids) after the runner's
|
|
// retry budget. The verifier is skipped for these steps so transient
|
|
// state does not poison the previous/current extractor advance.
|
|
Transitional bool `json:"transitional,omitempty"`
|
|
// SkippedVerification is set true exactly when the verifier was skipped
|
|
// for this step, so downstream tooling can tell a deliberately-skipped
|
|
// step from one that was verified and came back clean.
|
|
SkippedVerification bool `json:"skipped_verification,omitempty"`
|
|
// Witnesses records the violation witness for each property that newly
|
|
// violated at this step: the cause and the extractor values at onset.
|
|
Witnesses map[string]Witness `json:"witnesses,omitempty"`
|
|
}
|
|
|
|
// Witness is the trace-side record of a property violation: why it fired, the
|
|
// two steps a deferred obligation spans, and the extractor values behind it.
|
|
type Witness struct {
|
|
Reason string `json:"reason,omitempty"`
|
|
IsError bool `json:"is_error,omitempty"`
|
|
// Step is the step the failed obligation originated at: the step that
|
|
// armed it. For a deferred obligation (a next, an eventually) this is
|
|
// earlier than the step at which the failure was detected.
|
|
Step int `json:"step,omitempty"`
|
|
// DetectedStep is the observation whose evaluation produced the violation.
|
|
// Extractors is that step's state, not Step's.
|
|
DetectedStep int `json:"detected_step,omitempty"`
|
|
Extractors map[string]json.RawMessage `json:"extractors,omitempty"`
|
|
}
|
|
|
|
// ExtractorChange records the prev/curr JSON values of an extractor whose
|
|
// observation differed between two consecutive steps. Surfaced under
|
|
// violation rows in the replay UI as a "what changed at this step"
|
|
// breadcrumb.
|
|
type ExtractorChange struct {
|
|
Prev json.RawMessage `json:"prev"`
|
|
Curr json.RawMessage `json:"curr"`
|
|
}
|
|
|
|
type Metrics struct {
|
|
CPUPercent float64 `json:"cpu_percent"`
|
|
HeapBytes int64 `json:"heap_bytes,omitempty"`
|
|
TotalMemoryBytes int64 `json:"total_memory_bytes,omitempty"`
|
|
}
|
|
|
|
type Action struct {
|
|
Kind string `json:"kind"`
|
|
X int `json:"x,omitempty"`
|
|
Y int `json:"y,omitempty"`
|
|
FromX int `json:"from_x,omitempty"`
|
|
FromY int `json:"from_y,omitempty"`
|
|
ToX int `json:"to_x,omitempty"`
|
|
ToY int `json:"to_y,omitempty"`
|
|
Key string `json:"key,omitempty"`
|
|
Text string `json:"text,omitempty"`
|
|
DurationMillis int `json:"duration_millis,omitempty"`
|
|
Selector string `json:"selector,omitempty"`
|
|
ResolvedBounds *BoundsRecord `json:"resolved_bounds,omitempty"`
|
|
TapPoint *PointRecord `json:"tap_point,omitempty"`
|
|
// Source names the backend that chose this action: "llm" when the LLM
|
|
// action backend selected it, empty for the seeded picker. LLMReasoning is
|
|
// the model's short rationale, shown by the replay UI to explain the pick.
|
|
Source string `json:"source,omitempty"`
|
|
LLMReasoning string `json:"llm_reasoning,omitempty"`
|
|
// LLMChoice is the 1-based number the model picked from the candidate list;
|
|
// LLMChosenAction is the action description it echoed for that number. The
|
|
// runner strict-skips when the echo disagrees with the numbered entry, so on
|
|
// a recorded action the two always agree; the replay UI shows them to
|
|
// confirm the reasoning matched the executed action.
|
|
LLMChoice int `json:"llm_choice,omitempty"`
|
|
LLMChosenAction string `json:"llm_chosen_action,omitempty"`
|
|
}
|
|
|
|
type BoundsRecord struct {
|
|
X int `json:"x"`
|
|
Y int `json:"y"`
|
|
Width int `json:"width"`
|
|
Height int `json:"height"`
|
|
}
|
|
|
|
type PointRecord struct {
|
|
X int `json:"x"`
|
|
Y int `json:"y"`
|
|
}
|
|
|
|
type Exception struct {
|
|
Class string `json:"class"`
|
|
Message string `json:"message,omitempty"`
|
|
StackTrace string `json:"stack_trace,omitempty"`
|
|
UnixMillis int64 `json:"unix_millis,omitempty"`
|
|
}
|
|
|
|
type Meta struct {
|
|
Seed int64 `json:"seed"`
|
|
SpecPath string `json:"spec_path"`
|
|
BundleSHA256 string `json:"bundle_sha256"`
|
|
Platform string `json:"platform"`
|
|
BundleID string `json:"bundle_id"`
|
|
StartedAt time.Time `json:"started_at"`
|
|
EndedAt *time.Time `json:"ended_at,omitempty"`
|
|
SanderlingVersion string `json:"sanderling_version"`
|
|
|
|
// Arm labels the experiment cell this run belongs to, set from --arm. A
|
|
// directory of runs cannot be attributed to a cell after the fact without
|
|
// it, which makes any factorial computed from such a directory unanalysable.
|
|
Arm string `json:"arm,omitempty"`
|
|
// Generator, Model and Instructions record which picker ran and how it was
|
|
// configured. Two runs that differ in any of these are different arms.
|
|
Generator string `json:"generator,omitempty"`
|
|
Model string `json:"model,omitempty"`
|
|
Instructions string `json:"instructions,omitempty"`
|
|
// MaxSteps and DurationMillis are the budget the run was given, which has
|
|
// to be identical across arms for a comparison to mean anything.
|
|
MaxSteps int `json:"max_steps,omitempty"`
|
|
DurationMillis int64 `json:"duration_millis,omitempty"`
|
|
// Host is the machine that produced the run. Campaigns are split across
|
|
// several hosts, so a per-host effect has to be detectable rather than
|
|
// invisible.
|
|
Host string `json:"host,omitempty"`
|
|
}
|
|
|
|
type Writer struct {
|
|
directory string
|
|
mutex sync.Mutex
|
|
file io.WriteCloser
|
|
encoder *json.Encoder
|
|
}
|
|
|
|
// NewWriter ensures `directory` exists and opens trace.jsonl for append.
|
|
// meta.json is written separately via WriteMeta. Caller must Close.
|
|
func NewWriter(directory string) (*Writer, error) {
|
|
if err := os.MkdirAll(directory, 0o755); err != nil {
|
|
return nil, fmt.Errorf("mkdir: %w", err)
|
|
}
|
|
file, err := os.OpenFile(
|
|
filepath.Join(directory, "trace.jsonl"),
|
|
os.O_CREATE|os.O_WRONLY|os.O_APPEND,
|
|
0o644,
|
|
)
|
|
if err != nil {
|
|
return nil, fmt.Errorf("open trace.jsonl: %w", err)
|
|
}
|
|
encoder := json.NewEncoder(file)
|
|
return &Writer{directory: directory, file: file, encoder: encoder}, nil
|
|
}
|
|
|
|
func (w *Writer) Directory() string { return w.directory }
|
|
|
|
func (w *Writer) WriteMeta(meta Meta) error {
|
|
body, err := json.MarshalIndent(meta, "", " ")
|
|
if err != nil {
|
|
return fmt.Errorf("marshal meta: %w", err)
|
|
}
|
|
return os.WriteFile(filepath.Join(w.directory, "meta.json"), body, 0o644)
|
|
}
|
|
|
|
func (w *Writer) WriteStep(step Step) error {
|
|
w.mutex.Lock()
|
|
defer w.mutex.Unlock()
|
|
if w.file == nil {
|
|
return fmt.Errorf("trace: writer is closed")
|
|
}
|
|
return w.encoder.Encode(step)
|
|
}
|
|
|
|
// WriteScreenshot is lock-free: each call writes a distinct, uniquely-named
|
|
// file via os.WriteFile and touches no field of Writer, so concurrent calls
|
|
// never contend.
|
|
func (w *Writer) WriteScreenshot(stepIndex int, png []byte) error {
|
|
return w.writePNG(fmt.Sprintf("step-%05d.png", stepIndex), png)
|
|
}
|
|
|
|
func (w *Writer) writePNG(name string, png []byte) error {
|
|
if len(png) == 0 {
|
|
return nil
|
|
}
|
|
directory := filepath.Join(w.directory, "screenshots")
|
|
if err := os.MkdirAll(directory, 0o755); err != nil {
|
|
return fmt.Errorf("mkdir screenshots: %w", err)
|
|
}
|
|
return os.WriteFile(filepath.Join(directory, name), png, 0o644)
|
|
}
|
|
|
|
func (w *Writer) Close() error {
|
|
w.mutex.Lock()
|
|
defer w.mutex.Unlock()
|
|
if w.file == nil {
|
|
return nil
|
|
}
|
|
err := w.file.Close()
|
|
w.file = nil
|
|
return err
|
|
}
|