zoho: fix inefficiencies uploading with new API to avoid throttling

Before this fix, rclone queried the uploaded object to find its size
and modtime after upload as the API did not return these items.

Zoho have subsequently modified the API to return these items so
rclone uses them to avoid an API call.

This should help with rclone being throttled by Zoho.

See: https://forum.rclone.org/t/second-followup-on-the-older-topic-rclone-invokes-more-number-of-workdrive-s-files-listing-api-calls-which-exceeds-the-throttling-limit/45697/20
This commit is contained in:
Nick Craig-Wood 2024-08-27 09:06:36 +01:00
parent 9deb3e8adf
commit 802a938bd1
2 changed files with 105 additions and 18 deletions

View File

@ -2,6 +2,8 @@
package api
import (
"encoding/json"
"fmt"
"strconv"
"time"
)
@ -12,7 +14,12 @@ type Time time.Time
// UnmarshalJSON turns JSON into a Time
func (t *Time) UnmarshalJSON(data []byte) error {
millis, err := strconv.ParseInt(string(data), 10, 64)
s := string(data)
// If the time is a quoted string, strip quotes
if len(s) >= 2 && s[0] == '"' && s[len(s)-1] == '"' {
s = s[1 : len(s)-1]
}
millis, err := strconv.ParseInt(s, 10, 64)
if err != nil {
return err
}
@ -84,6 +91,73 @@ type ItemList struct {
Items []Item `json:"data"`
}
// UploadFileInfo is what the FileInfo field in the UnloadInfo struct decodes to
type UploadFileInfo struct {
OrgID string `json:"ORG_ID"`
ResourceID string `json:"RESOURCE_ID"`
LibraryID string `json:"LIBRARY_ID"`
Md5Checksum string `json:"MD5_CHECKSUM"`
ParentModelID string `json:"PARENT_MODEL_ID"`
ParentID string `json:"PARENT_ID"`
ResourceType int `json:"RESOURCE_TYPE"`
WmsSentTime string `json:"WMS_SENT_TIME"`
TabID string `json:"TAB_ID"`
Owner string `json:"OWNER"`
ResourceGroup string `json:"RESOURCE_GROUP"`
ParentModelName string `json:"PARENT_MODEL_NAME"`
Size int64 `json:"size"`
Operation string `json:"OPERATION"`
EventID string `json:"EVENT_ID"`
AuditInfo struct {
VersionInfo struct {
VersionAuthors []string `json:"versionAuthors"`
VersionID string `json:"versionId"`
IsMinorVersion bool `json:"isMinorVersion"`
VersionTime Time `json:"versionTime"`
VersionAuthorZuid []string `json:"versionAuthorZuid"`
VersionNotes string `json:"versionNotes"`
VersionNumber string `json:"versionNumber"`
} `json:"versionInfo"`
Resource struct {
Owner string `json:"owner"`
CreatedTime Time `json:"created_time"`
Creator string `json:"creator"`
ServiceType int `json:"service_type"`
Extension string `json:"extension"`
StatusChangeTime Time `json:"status_change_time"`
ResourceType int `json:"resource_type"`
Name string `json:"name"`
} `json:"resource"`
ParentInfo struct {
ParentName string `json:"parentName"`
ParentID string `json:"parentId"`
ParentType int `json:"parentType"`
} `json:"parentInfo"`
LibraryInfo struct {
LibraryName string `json:"libraryName"`
LibraryID string `json:"libraryId"`
LibraryType int `json:"libraryType"`
} `json:"libraryInfo"`
UpdateType string `json:"updateType"`
StatusCode string `json:"statusCode"`
} `json:"AUDIT_INFO"`
ZUID int64 `json:"ZUID"`
TeamID string `json:"TEAM_ID"`
}
// GetModTime fetches the modification time of the upload
//
// This tries a few places and if all fails returns the current time
func (ufi *UploadFileInfo) GetModTime() Time {
if t := ufi.AuditInfo.Resource.CreatedTime; !time.Time(t).IsZero() {
return t
}
if t := ufi.AuditInfo.Resource.StatusChangeTime; !time.Time(t).IsZero() {
return t
}
return Time(time.Now())
}
// UploadInfo is a simplified and slightly different version of
// the Item struct only used in the response to uploads
type UploadInfo struct {
@ -91,9 +165,21 @@ type UploadInfo struct {
ParentID string `json:"parent_id"`
FileName string `json:"notes.txt"`
RessourceID string `json:"resource_id"`
Permalink string `json:"Permalink"`
FileInfo string `json:"File INFO"` // JSON encoded UploadFileInfo
} `json:"attributes"`
}
// GetUploadFileInfo decodes the embedded FileInfo
func (ui *UploadInfo) GetUploadFileInfo() (*UploadFileInfo, error) {
var ufi UploadFileInfo
err := json.Unmarshal([]byte(ui.Attributes.FileInfo), &ufi)
if err != nil {
return nil, fmt.Errorf("failed to decode FileInfo: %w", err)
}
return &ufi, nil
}
// UploadResponse is the response to a file Upload
type UploadResponse struct {
Uploads []UploadInfo `json:"data"`

View File

@ -677,25 +677,26 @@ func (f *Fs) upload(ctx context.Context, name string, parent string, size int64,
if len(uploadResponse.Uploads) != 1 {
return nil, errors.New("upload: invalid response")
}
// Received meta data is missing size so we have to read it again.
// It doesn't always appear on first read so try again if necessary
var info *api.Item
const maxTries = 10
sleepTime := 100 * time.Millisecond
for i := 0; i < maxTries; i++ {
info, err = f.readMetaDataForID(ctx, uploadResponse.Uploads[0].Attributes.RessourceID)
if err != nil {
return nil, err
}
if info.Attributes.StorageInfo.Size != 0 || size == 0 {
break
}
fs.Debugf(f, "Size not available yet for %q - try again in %v (try %d/%d)", name, sleepTime, i+1, maxTries)
time.Sleep(sleepTime)
sleepTime *= 2
upload := uploadResponse.Uploads[0]
uploadInfo, err := upload.GetUploadFileInfo()
if err != nil {
return nil, fmt.Errorf("upload error: %w", err)
}
return info, nil
// Fill in the api.Item from the api.UploadFileInfo
var info api.Item
info.ID = upload.Attributes.RessourceID
info.Attributes.Name = upload.Attributes.FileName
// info.Attributes.Type = not used
info.Attributes.IsFolder = false
// info.Attributes.CreatedTime = not used
info.Attributes.ModifiedTime = uploadInfo.GetModTime()
// info.Attributes.UploadedTime = 0 not used
info.Attributes.StorageInfo.Size = uploadInfo.Size
info.Attributes.StorageInfo.FileCount = 0
info.Attributes.StorageInfo.FolderCount = 0
return &info, nil
}
// Put the object into the container