Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Always numerize the result path of containerized runs to avoid I/O sync problems #274

Merged
merged 1 commit into from
Jul 18, 2024
Merged

Always numerize the result path of containerized runs to avoid I/O sync problems #274

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 20 additions & 6 deletions cmd/eval-dev-quality/cmd/evaluate.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,7 @@ import (
"path/filepath"
"slices"
"sort"
"strconv"
"strings"
"text/template"
"time"
Expand Down Expand Up @@ -503,7 +504,8 @@ func (command *Evaluate) evaluateDocker(ctx *evaluate.Context) (err error) {
}

// Iterate over each model and start the container.
for _, model := range ctx.Models {
models := map[string]bool{}
for i, model := range ctx.Models {
// We are skipping ollama models until we fully support pulling. https://github.com/symflower/eval-dev-quality/issues/100.
if ctx.ProviderForModel[model].ID() == "ollama" {
command.logger.Print("Skipping unsupported ollama model with docker runtime")
Expand All @@ -522,11 +524,18 @@ func (command *Evaluate) evaluateDocker(ctx *evaluate.Context) (err error) {
}

// Commands for the evaluation to run inside the container.
resultPath := "/app/evaluation/" + log.CleanModelNameForFileSystem(model.ID())
if models[model.ID()] {
resultPath += "_" + strconv.Itoa(i)
} else {
models[model.ID()] = true
}

evaluationCommand := []string{
"eval-dev-quality",
"evaluate",
"--model", model.ID(),
"--result-path", "/app/evaluation/" + model.ID(),
"--result-path", resultPath,
}

cmd := append(dockerCommand, evaluationCommand...)
Expand Down Expand Up @@ -618,6 +627,7 @@ func (command *Evaluate) evaluateKubernetes(ctx *evaluate.Context) (err error) {
parallel := util.NewParallel(command.Parallel)

// Iterate over each model and start the container.
models := map[string]bool{}
for i, model := range ctx.Models {
// We are skipping ollama models until we fully support pulling. https://github.com/symflower/eval-dev-quality/issues/100.
if ctx.ProviderForModel[model].ID() == "ollama" {
Expand All @@ -635,13 +645,17 @@ func (command *Evaluate) evaluateKubernetes(ctx *evaluate.Context) (err error) {
}

// Commands for the evaluation to run inside the container.
resultPath := "/var/evaluation/" + log.CleanModelNameForFileSystem(model.ID())
if models[model.ID()] {
resultPath += "_" + strconv.Itoa(i)
} else {
models[model.ID()] = true
}
evaluationCommand := []string{
"eval-dev-quality",
"evaluate",
"--model",
model.ID(),
"--result-path",
"/var/evaluations/" + model.ID(),
"--model", model.ID(),
"--result-path", resultPath,
}
cmd := append(evaluationCommand, args...)

Expand Down
36 changes: 18 additions & 18 deletions cmd/eval-dev-quality/cmd/evaluate_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -809,8 +809,8 @@ func TestEvaluateExecute(t *testing.T) {
filepath.Join("result-directory", "evaluation.log"): nil,

// Parallel run 1
filepath.Join("result-directory", "symflower", "symbolic-execution", "categories.svg"): nil,
filepath.Join("result-directory", "symflower", "symbolic-execution", "evaluation.csv"): func(t *testing.T, filePath, data string) {
filepath.Join("result-directory", "symflower_symbolic-execution", "categories.svg"): nil,
filepath.Join("result-directory", "symflower_symbolic-execution", "evaluation.csv"): func(t *testing.T, filePath, data string) {
actualAssessments := validateMetrics(t, extractMetricsCSVMatch, data, []metrics.Assessments{
metrics.Assessments{
metrics.AssessmentKeyCoverage: 10,
Expand Down Expand Up @@ -850,18 +850,18 @@ func TestEvaluateExecute(t *testing.T) {
assert.Equal(t, uint64(254), actualAssessments[0][metrics.AssessmentKeyGenerateTestsForFileCharacterCount])
assert.Equal(t, uint64(254), actualAssessments[0][metrics.AssessmentKeyResponseCharacterCount])
},
filepath.Join("result-directory", "symflower", "symbolic-execution", "evaluation.log"): nil,
filepath.Join("result-directory", "symflower", "symbolic-execution", "README.md"): nil,
filepath.Join("result-directory", "symflower", "symbolic-execution", string(evaluatetask.IdentifierWriteTests), "symflower_symbolic-execution", "golang", "golang", "plain.log"): func(t *testing.T, filePath, data string) {
filepath.Join("result-directory", "symflower_symbolic-execution", "evaluation.log"): nil,
filepath.Join("result-directory", "symflower_symbolic-execution", "README.md"): nil,
filepath.Join("result-directory", "symflower_symbolic-execution", string(evaluatetask.IdentifierWriteTests), "symflower_symbolic-execution", "golang", "golang", "plain.log"): func(t *testing.T, filePath, data string) {
assert.Equal(t, 1, strings.Count(data, `Evaluating model "symflower/symbolic-execution"`))
},
filepath.Join("result-directory", "symflower", "symbolic-execution", string(evaluatetask.IdentifierWriteTests), "symflower_symbolic-execution", "java", "java", "plain.log"): func(t *testing.T, filePath, data string) {
filepath.Join("result-directory", "symflower_symbolic-execution", string(evaluatetask.IdentifierWriteTests), "symflower_symbolic-execution", "java", "java", "plain.log"): func(t *testing.T, filePath, data string) {
assert.Equal(t, 1, strings.Count(data, `Evaluating model "symflower/symbolic-execution"`))
},

// Parallel run 2
filepath.Join("result-directory", "symflower", "symbolic-execution-0", "categories.svg"): nil,
filepath.Join("result-directory", "symflower", "symbolic-execution-0", "evaluation.csv"): func(t *testing.T, filePath, data string) {
filepath.Join("result-directory", "symflower_symbolic-execution_1", "categories.svg"): nil,
filepath.Join("result-directory", "symflower_symbolic-execution_1", "evaluation.csv"): func(t *testing.T, filePath, data string) {
actualAssessments := validateMetrics(t, extractMetricsCSVMatch, data, []metrics.Assessments{
metrics.Assessments{
metrics.AssessmentKeyCoverage: 10,
Expand Down Expand Up @@ -901,18 +901,18 @@ func TestEvaluateExecute(t *testing.T) {
assert.Equal(t, uint64(254), actualAssessments[0][metrics.AssessmentKeyGenerateTestsForFileCharacterCount])
assert.Equal(t, uint64(254), actualAssessments[0][metrics.AssessmentKeyResponseCharacterCount])
},
filepath.Join("result-directory", "symflower", "symbolic-execution-0", "evaluation.log"): nil,
filepath.Join("result-directory", "symflower", "symbolic-execution-0", "README.md"): nil,
filepath.Join("result-directory", "symflower", "symbolic-execution-0", string(evaluatetask.IdentifierWriteTests), "symflower_symbolic-execution", "golang", "golang", "plain.log"): func(t *testing.T, filePath, data string) {
filepath.Join("result-directory", "symflower_symbolic-execution_1", "evaluation.log"): nil,
filepath.Join("result-directory", "symflower_symbolic-execution_1", "README.md"): nil,
filepath.Join("result-directory", "symflower_symbolic-execution_1", string(evaluatetask.IdentifierWriteTests), "symflower_symbolic-execution", "golang", "golang", "plain.log"): func(t *testing.T, filePath, data string) {
assert.Equal(t, 1, strings.Count(data, `Evaluating model "symflower/symbolic-execution"`))
},
filepath.Join("result-directory", "symflower", "symbolic-execution-0", string(evaluatetask.IdentifierWriteTests), "symflower_symbolic-execution", "java", "java", "plain.log"): func(t *testing.T, filePath, data string) {
filepath.Join("result-directory", "symflower_symbolic-execution_1", string(evaluatetask.IdentifierWriteTests), "symflower_symbolic-execution", "java", "java", "plain.log"): func(t *testing.T, filePath, data string) {
assert.Equal(t, 1, strings.Count(data, `Evaluating model "symflower/symbolic-execution"`))
},

// Parallel run 3
filepath.Join("result-directory", "symflower", "symbolic-execution-1", "categories.svg"): nil,
filepath.Join("result-directory", "symflower", "symbolic-execution-1", "evaluation.csv"): func(t *testing.T, filePath, data string) {
filepath.Join("result-directory", "symflower_symbolic-execution_2", "categories.svg"): nil,
filepath.Join("result-directory", "symflower_symbolic-execution_2", "evaluation.csv"): func(t *testing.T, filePath, data string) {
actualAssessments := validateMetrics(t, extractMetricsCSVMatch, data, []metrics.Assessments{
metrics.Assessments{
metrics.AssessmentKeyCoverage: 10,
Expand Down Expand Up @@ -952,12 +952,12 @@ func TestEvaluateExecute(t *testing.T) {
assert.Equal(t, uint64(254), actualAssessments[0][metrics.AssessmentKeyGenerateTestsForFileCharacterCount])
assert.Equal(t, uint64(254), actualAssessments[0][metrics.AssessmentKeyResponseCharacterCount])
},
filepath.Join("result-directory", "symflower", "symbolic-execution-1", "evaluation.log"): nil,
filepath.Join("result-directory", "symflower", "symbolic-execution-1", "README.md"): nil,
filepath.Join("result-directory", "symflower", "symbolic-execution-1", string(evaluatetask.IdentifierWriteTests), "symflower_symbolic-execution", "golang", "golang", "plain.log"): func(t *testing.T, filePath, data string) {
filepath.Join("result-directory", "symflower_symbolic-execution_2", "evaluation.log"): nil,
filepath.Join("result-directory", "symflower_symbolic-execution_2", "README.md"): nil,
filepath.Join("result-directory", "symflower_symbolic-execution_2", string(evaluatetask.IdentifierWriteTests), "symflower_symbolic-execution", "golang", "golang", "plain.log"): func(t *testing.T, filePath, data string) {
assert.Equal(t, 1, strings.Count(data, `Evaluating model "symflower/symbolic-execution"`))
},
filepath.Join("result-directory", "symflower", "symbolic-execution-1", string(evaluatetask.IdentifierWriteTests), "symflower_symbolic-execution", "java", "java", "plain.log"): func(t *testing.T, filePath, data string) {
filepath.Join("result-directory", "symflower_symbolic-execution_2", string(evaluatetask.IdentifierWriteTests), "symflower_symbolic-execution", "java", "java", "plain.log"): func(t *testing.T, filePath, data string) {
assert.Equal(t, 1, strings.Count(data, `Evaluating model "symflower/symbolic-execution"`))
},
},
Expand Down