diff options
author | Martin Czygan <martin.czygan@gmail.com> | 2021-03-31 23:50:52 +0200 |
---|---|---|
committer | Martin Czygan <martin.czygan@gmail.com> | 2021-03-31 23:50:52 +0200 |
commit | a7e0e8fe3a78fa6a9872cefcb7a9eb261128f6f2 (patch) | |
tree | 2850c2ed0ecd0392deab2fe82c30c78465f1576c /skate/cmd | |
parent | cedef349990c12747efdbe10f692c57436dc5c91 (diff) | |
download | refcat-a7e0e8fe3a78fa6a9872cefcb7a9eb261128f6f2.tar.gz refcat-a7e0e8fe3a78fa6a9872cefcb7a9eb261128f6f2.zip |
cleanup command
Diffstat (limited to 'skate/cmd')
-rw-r--r-- | skate/cmd/skate-cluster-stats/main.go | 92 |
1 files changed, 0 insertions, 92 deletions
diff --git a/skate/cmd/skate-cluster-stats/main.go b/skate/cmd/skate-cluster-stats/main.go deleted file mode 100644 index 4973b4d..0000000 --- a/skate/cmd/skate-cluster-stats/main.go +++ /dev/null @@ -1,92 +0,0 @@ -package main - -import ( - "flag" - "fmt" - "log" - "os" - "runtime" - - jsoniter "github.com/json-iterator/go" - "git.archive.org/martin/cgraph/skate" - "git.archive.org/martin/cgraph/skate/parallel" -) - -var ( - numWorkers = flag.Int("w", runtime.NumCPU(), "number of workers") - batchSize = flag.Int("b", 100000, "batch size") - bestEffort = flag.Bool("B", false, "best effort, log errors") - // unmatched: clusters w/ refs only - // count: number of entities in cluster (by type) - // default: key and number of values - mode = flag.String("m", "", "what to extract (unmatched, count, ...)") - - json = jsoniter.ConfigCompatibleWithStandardLibrary - bytesNewline = []byte("\n") -) - -type Func func([]byte) ([]byte, error) - -func main() { - flag.Parse() - var f Func - switch *mode { - case "unmatched": - f = func(p []byte) ([]byte, error) { - var cluster skate.ReleaseCluster - if err := json.Unmarshal(p, &cluster); err != nil { - if *bestEffort { - log.Printf("%v", err) - return nil, nil - } - log.Fatal(err) - } - var refs int - for _, v := range cluster.Values { - if v.Extra.Skate.Status == "ref" { - refs++ - } - } - if refs == len(cluster.Values) { - return p, nil - } - return nil, nil - } - case "count": - f = func(p []byte) ([]byte, error) { - var cluster skate.ReleaseCluster - if err := json.Unmarshal(p, &cluster); err != nil { - if *bestEffort { - log.Printf("%v", err) - return nil, nil - } - log.Fatal(err) - } - var refs int - for _, v := range cluster.Values { - if v.Extra.Skate.Status == "ref" { - refs++ - } - } - // total, refs, non-refs, key - s := fmt.Sprintf("%d\t%d\t%d\t%s\n", - len(cluster.Values), refs, len(cluster.Values)-refs, cluster.Key) - return []byte(s), nil - } - default: - f = func(p []byte) ([]byte, error) { - var cluster skate.ReleaseCluster - if err := json.Unmarshal(p, &cluster); err != nil { - return nil, err - } - s := fmt.Sprintf("%d\t%s\n", len(cluster.Values), cluster.Key) - return []byte(s), nil - } - } - pp := parallel.NewProcessor(os.Stdin, os.Stdout, f) - pp.NumWorkers = *numWorkers - pp.BatchSize = *batchSize - if err := pp.Run(); err != nil { - log.Fatal(err) - } -} |