Use new data-stream-list server script data structure

- add to report method saves new large-hash parameter
- delete from report updated
- split download updated
- resume operation (both full and split files) updated
- warn conflicts updated
- update dataset operation updated
- rename operation (download and ingest)
- download discard changes updated
- stage feature works properly (add/remove local changes)
- ingestion reportUpToDate updated
- split ingestion updated
- dataset utils remove hardcoded url
- report up to date method updated

ebulk-data/embulk-wendelin-dataset-tool/lib/embulk/dataset_utils.rb

@@ -6,18 +6,9 @@ require 'open-uri'
 # class that handles dataset tasks report
 class DatasetUtils
 
-  begin
-    uri = URI(URI.escape("https://wendelin.io/ERP5Site_getIngestionConstantsJson"))
-    res = open(uri).read
-    json = JSON.parse(res)
-    EOF = json['split_end_suffix']
-    NONE_EXT = json['none_extension']
-    REFERENCE_SEPARATOR = json['reference_separator']
-  rescue
-    EOF = "EOF"
-    NONE_EXT = "none"
-    REFERENCE_SEPARATOR = "/"
-  end
+  EOF = "EOF"
+  NONE_EXT = "none"
+  REFERENCE_SEPARATOR = "/"
 
   DATASET_REPORT_FILE = ".dataset-task-report"
   DATASET_TEMP_REPORT_FILE = ".temp-dataset-task-report"
@@ -68,7 +59,21 @@ class DatasetUtils
   RECORD_SEPARATOR = ";"
   DATE_FORMAT = "%Y-%m-%d-%H-%M-%S"
 
-  def initialize(data_set_directory)
+  def initialize(data_set_directory, url=FALSE)
+    if url
+      begin
+        uri = URI(URI.escape(url + "/ERP5Site_getIngestionConstantsJson"))
+        res = open(uri).read
+        json = JSON.parse(res)
+        @EOF = json['split_end_suffix']
+        @NONE_EXT = json['none_extension']
+        @REFERENCE_SEPARATOR = json['reference_separator']
+      rescue
+        @EOF = EOF
+        @NONE_EXT = NONE_EXT
+        @REFERENCE_SEPARATOR = REFERENCE_SEPARATOR
+      end
+    end
     @data_set_directory = appendSlashTo(data_set_directory)
     @logger = LogManager.instance()
     @task_report_file = @data_set_directory + DATASET_REPORT_FILE
@@ -179,7 +184,7 @@ class DatasetUtils
 	if record[1].chomp == RUN_DONE
 	  if (remove.nil?) || (remove != record[0])
 	    local_files[record[0]] = {"size" => record[2].chomp, "hash" => record[3].chomp,
-				      "status" => record[1].chomp, "modification_date" => record[4].chomp }
+				      "status" => record[1].chomp, "modification_date" => record[4].chomp, "large_hash" => record[5].chomp }
 	  end
 	end
       end
@@ -197,7 +202,7 @@ class DatasetUtils
 	File.open(@temp_report_file, 'w') {}
       else
 	local_files.each do |key, array|
-	  record = [key, array["status"], array["size"].to_s, array["hash"], array["modification_date"]].join(RECORD_SEPARATOR)
+	  record = [key, array["status"], array["size"].to_s, array["hash"], array["modification_date"], array["large_hash"]].join(RECORD_SEPARATOR)
 	  File.open(@temp_report_file, 'ab') { |file| file.puts(record) }
 	end
       end
@@ -262,9 +267,10 @@ class DatasetUtils
 	    else
 	      file_path = referenceToPath(changes[0]["reference"], @data_set_directory, data_set)
 	    end
-	    size = changes[0]["size"]
-	    hash = changes[0]["hash"]
-	    addToReport(changes[0]["reference"], RUN_DONE, size, hash, data_set, new_reference)
+	    size = changes[0]["full-size"]
+	    large_hash = changes[0]["large-hash"]
+	    hash = getHash(file_path).to_s
+	    addToReport(changes[0]["reference"], RUN_DONE, size, hash, large_hash, data_set, new_reference)
 	    File.delete(@resume_operation_file)
 	    return TRUE
 	  end
@@ -376,9 +382,9 @@ class DatasetUtils
     end
   end
 
-  def saveSplitOperation(operation, reference, eof, hash, chunk_size)
+  def saveSplitOperation(operation, reference, eof, hash, chunk_size, large_hash="NONE")
     file_reference = reference.gsub('/', '__')
-    record = [operation, reference, eof, hash, Integer(chunk_size)].join(RECORD_SEPARATOR)
+    record = [operation, reference, eof, hash, Integer(chunk_size), large_hash].join(RECORD_SEPARATOR)
     File.open(@split_file + "__" + file_reference, 'w') { |file| file.puts(record) }
     deleteSplitOperationControlFile(reference)
   end
@@ -388,9 +394,16 @@ class DatasetUtils
     return File.exist?(@split_file + "__" + file_reference)
   end
 
-  def getLastSplitOperation(operation, reference, hash, chunk_size)
+  def getLastSplitOperation(operation, reference, hash, chunk_size, large_hash=FALSE)
+    # returns last split operation index (and partial large hash if large_hash is TRUE)
     file_reference = reference.gsub('/', '__')
-    return 0 if not File.exist?(@split_file + "__" + file_reference)
+    if not File.exist?(@split_file + "__" + file_reference)
+      if large_hash
+        return 0, ""
+      else
+        return 0
+      end
+    end
     record = File.open(@split_file + "__" + file_reference).read.chomp.split(RECORD_SEPARATOR)
     if record[0] == operation && record[1] == reference && record[3] == hash && record[4] == Integer(chunk_size).to_s && record[2] != EOF
       # discard if user interrupted (ctrl+c) the operation
@@ -398,16 +411,32 @@ class DatasetUtils
 	@logger.warn("Previous split operation attempt for file #{reference} was interrupt by user (aborted tool execution), it will be restarted.", print=TRUE)
 	deleteSplitOperationFile(file_reference)
 	deleteSplitOperationControlFile(file_reference)
-	return 0
+	if large_hash
+    return 0, ""
+  else
+    return 0
+  end
       end
       createSplitOperationControlFile(file_reference)
-      return record[2].to_i
+      if large_hash
+        return record[2].to_i, record[5]
+      else
+        return record[2].to_i
+      end
+    end
+    if large_hash
+      return 0, ""
+    else
+      return 0
     end
-    return 0
   rescue Exception => e
     @logger.error("An error occurred in getLastSplitOperation method:" + e.to_s)
     @logger.error(e.backtrace)
-    return 0
+    if large_hash
+      return 0, ""
+    else
+      return 0
+    end
   end
 
   def deleteDiscardChangesFile()
@@ -502,7 +531,7 @@ class DatasetUtils
     return filename, extension, reference
   end
 
-  def addToReport(reference, status, size, hash, data_set, new_reference=FALSE)
+  def addToReport(reference, status, size, hash, large_hash, data_set, new_reference=FALSE)
     local_files = {}
     begin
       file_path = referenceToPath(reference, @data_set_directory, data_set)
@@ -518,18 +547,19 @@ class DatasetUtils
 	    reference = new_reference
 	    file_path = referenceToPath(reference, @data_set_directory, data_set)
 	    modification_date = File.exist?(file_path) ? File.mtime(file_path).strftime(DATE_FORMAT) : "not-modification-date"
+	    large_hash = record[5] if large_hash == ""
 	  end
 	  local_files[reference] = {"size" => size, "hash" => hash, "status" => status,
-				    "modification_date" => modification_date }
+				    "modification_date" => modification_date, "large_hash" => large_hash }
 	  new_file = FALSE
 	else
 	  local_files[record[0]] = {"size" => record[2].chomp, "hash" => record[3].chomp,
-				    "status" => record[1].chomp, "modification_date" => record[4].chomp }
+				    "status" => record[1].chomp, "modification_date" => record[4].chomp, "large_hash" => record[5] }
 	end
       end
       if new_file
 	local_files[reference] = {"size" => size, "hash" => hash, "status" => status,
-				  "modification_date" => modification_date }
+				  "modification_date" => modification_date, "large_hash" => large_hash }
       end
     rescue Exception => e
       @logger.error("An error occurred in DatasetUtils method 'addToReport':" + e.to_s)
@@ -543,9 +573,21 @@ class DatasetUtils
     saveReport(local_files)
   end
 
+  def getHashFromChunk(content)
+    begin
+      md5 = Digest::MD5.new
+      md5.update(content)
+      return md5.hexdigest
+    rescue Exception => e
+      @logger.error("An error occurred while getting hash from chunk:" + e.to_s, print=TRUE)
+      @logger.error(e.backtrace)
+      raise e
+    end
+  end
+
   def getHash(file)
-    return "FILE-NOT-EXISTS" if ! File.exist?(file)
     begin
+      raise "File doesn't exist" if ! File.exist?(file)
       chunk_size = 4 * MEGA
       md5 = Digest::MD5.new
       open(file) do |f|
@@ -608,16 +650,32 @@ class DatasetUtils
     end
   end
 
-  def isRenamed(file, file_dict_list, file_dict=FALSE)
-    hash = file_dict ? file_dict["hash"] : ""
-    size = file_dict ? file_dict["size"] : (File.size(file).to_s if File.exist?(file))
-    file_dict_list.each do |path, dict|
-      if size == dict["size"].to_s
-	hash = hash != "" ? hash : getHash(file).to_s
-	if hash == dict["hash"]
-	  old_path = path
-	  return {"key" => old_path, "size" => size, "hash" => hash}
-	end
+  def isLocalChangeRenamed(local_file_path, deleted_file_dict_list)
+    # checking if local_file_path marked as new is also a deleted file, so it's in fact a renamed
+    size = File.size(local_file_path).to_s if File.exist?(local_file_path)
+    deleted_file_dict_list.each do |path, deleted_file_dict|
+      if size == deleted_file_dict["size"].to_s
+        hash = getHash(local_file_path).to_s
+        if hash == deleted_file_dict["hash"]
+          old_path = path
+          return {"key" => old_path, "size" => size, "hash" => hash}
+        end
+      end
+    end
+    return FALSE
+  end
+
+  def isRemoteChangeRenamed(new_remote_file_dict_list, local_report_file_dict)
+    #checking if report local file missing in remote dataset is one of the remote files marked as new/modified, so it's in fact a remote rename
+    hash = local_report_file_dict["large-hash"]
+    size = local_report_file_dict["size"]
+    new_remote_file_dict_list.each do |path, new_remote_file_dict|
+      if size == new_remote_file_dict["full-size"].to_s
+        hash = hash != "" ? hash : getHash(file).to_s
+        if hash == new_remote_file_dict["large-hash"]
+          old_path = path
+          return {"key" => old_path, "size" => size, "hash" => hash}
+        end
       end
     end
     return FALSE
@@ -643,10 +701,10 @@ class DatasetUtils
     return staged_status["status"] == status
   end
 
-  def checkRenamed(path, deleted_files, change)
-    renamed_dict = isRenamed(path, deleted_files)
+  def checkRenamed(path, deleted_file_dict_list, change)
+    renamed_dict = isLocalChangeRenamed(path, deleted_file_dict_list)
     if renamed_dict
-      deleted_files.delete(renamed_dict["key"])
+      deleted_file_dict_list.delete(renamed_dict["key"])
       change["status"] = STATUS_RENAMED
       change["new_path"] = change["path"]
       change["path"] = renamed_dict["key"]
@@ -791,37 +849,37 @@ class DatasetUtils
 	File.readlines(@task_report_file).each do |line|
 	  record = line.split(RECORD_SEPARATOR)
 	  if record[1].chomp == RUN_DONE
-	    local_files[record[0]] = {"size" => record[2].chomp, "hash" => record[3].chomp, }
+	    local_files[record[0]] = {"size" => record[2].chomp, "hash" => record[3].chomp, "large-hash" => record[5].chomp, }
 	  end
 	end
 	data_streams.each do |data_stream|
-	  remote_files.push(data_stream["reference"])
-	  pending = TRUE
 	  reference = data_stream["reference"]
+	  remote_files.push(reference)
+	  pending = TRUE
 	  if local_files.has_key? reference
 	    size = local_files[reference]["size"]
-	    if size.to_s == data_stream["size"].to_s
-	      hash = local_files[reference]["hash"]
-	      if hash == data_stream["hash"] or data_stream["hash"] == ""
+	    if size.to_s == data_stream["full-size"].to_s
+	      hash = local_files[reference]["large-hash"]
+	      if hash == data_stream["large-hash"] or data_stream["large-hash"] == ""
 		pending = FALSE
 	      end
 	    end
 	  end
 	  if pending
 	    local_files.delete(reference)
-	    file_path = referenceToPath(data_stream["reference"], @data_set_directory, data_set)
+	    file_path = referenceToPath(reference, @data_set_directory, data_set)
 	    new_changed_files[file_path] = data_stream
 	  end
 	end
 	local_files.each do |reference, file_dict|
 	  if not remote_files.include? reference
-	    changed_data_stream = {"reference" => reference, "hash" => DELETE }
+	    changed_data_stream = {"reference" => reference, "large-hash" => DELETE }
 	    file_path = referenceToPath(reference, @data_set_directory, data_set)
-	    renamed_dict = isRenamed(file_path, new_changed_files, file_dict)
+	    renamed_dict = isRemoteChangeRenamed(new_changed_files, file_dict)
 	    if renamed_dict
 	      changed_data_stream = {"reference" => reference, "id" => new_changed_files[renamed_dict["key"]]["id"],
 				     "new_reference" => new_changed_files[renamed_dict["key"]]["reference"], "status" => STATUS_RENAMED,
-				     "size" => renamed_dict["size"], "hash" => renamed_dict["hash"] }
+				     "full-size" => renamed_dict["size"], "large-hash" => renamed_dict["hash"] }
 	      new_changed_files.delete(renamed_dict["key"])
 	    end
 	    changed_data_streams.push(changed_data_stream)

ebulk-data/embulk-wendelin-dataset-tool/lib/embulk/input/fif.rb

@@ -138,9 +138,9 @@ module Embulk
 		@logger.error("Your current dataset is outdated. Please, run a download to update it before ingest your changes.", print=TRUE)
 		puts
 		@logger.abortExecution(error=FALSE)
-	      end
-	    end
-	  end
+        end
+      end
+    end
 
 	  @logger.info("Supplier: #{task['supplier']}")
 	  @logger.info("Dataset name: #{task['data_set']}")
@@ -215,7 +215,7 @@ module Embulk
 	super
 	@supplier = task['supplier']
 	@dataset = task['data_set']
-	@chunk_size = task['chunk_size']
+	@chunk_size = DatasetUtils::CHUNK_SIZE
 	@data_set_directory = task['data_set_directory']
 	@logger = LogManager.instance()
 	@dataset_utils = DatasetUtils.new(@data_set_directory)
@@ -239,13 +239,19 @@ module Embulk
 	  filename, extension, reference = @dataset_utils.getPathInfo(path, @dataset)
 	  operation = rename ? DatasetUtils::RENAME : DatasetUtils::INGESTION
 	  @dataset_utils.saveCurrentOperation(operation, reference, new_reference)
-	  resume_split = @dataset_utils.splitOperationFileExist(reference) ? @dataset_utils.getLastSplitOperation(operation, reference, hash, @chunk_size) : 0
+	  resume_split, large_hash = 0, ""
+	  if @dataset_utils.splitOperationFileExist(reference)
+	    resume_split, large_hash = @dataset_utils.getLastSplitOperation(operation, reference, hash, @chunk_size, large_hash=TRUE)
+    end
 	  each_chunk(path, filename, extension, size, hash, schema[1..-1].map{|elm| elm.name}, @chunk_size, delete, new_reference, resume_split) do |entry|
 	    @dataset_utils.createSplitOperationControlFile(reference) if split
+	    large_hash += entry[8]
+	    #no need to send large hash to server
+	    entry.pop()
 	    @page_builder.add(entry)
 	    if ! delete && ! rename && entry[5] != ""
 	      split = TRUE
-	      @dataset_utils.saveSplitOperation(operation, reference, entry[5], hash, @chunk_size)
+	      @dataset_utils.saveSplitOperation(operation, reference, entry[5], hash, @chunk_size, large_hash)
 	    end
 	  end
 	  @page_builder.finish
@@ -269,7 +275,7 @@ module Embulk
 	    end
 	  else
 	    if @dataset_utils.reportFileExist()
-	      @dataset_utils.addToReport(reference, return_value, size, hash, task['data_set'], new_reference)
+	      @dataset_utils.addToReport(reference, return_value, size, hash, large_hash, task['data_set'], new_reference)
 	    end
 	  end
 	end
@@ -282,11 +288,11 @@ module Embulk
       def each_chunk(path, filename, extension, size, hash, fields, chunk_size=DatasetUtils::CHUNK_SIZE, delete=FALSE, new_reference=FALSE, resume_split=0)
 	if delete
 	  File.delete(path) if File.exist?(path)
-	  values = [@supplier, @dataset, filename, extension, "", DatasetUtils::DELETE, "", ""]
+	  values = [@supplier, @dataset, filename, extension, "", DatasetUtils::DELETE, "", "", ""]
 	  yield(values)
 	elsif new_reference
 	  File.delete(path) if File.exist?(path)
-	  values = [@supplier, @dataset, filename, extension, new_reference, DatasetUtils::RENAME, "", ""]
+	  values = [@supplier, @dataset, filename, extension, new_reference, DatasetUtils::RENAME, "", "", ""]
 	  yield(values)
 	else
 	  file_object = File.open(path, "rb")
@@ -297,7 +303,7 @@ module Embulk
 	      data = next_byte
 	      if not next_byte
 		if first # this means this is an empty file
-		  values = [@supplier, @dataset, filename, extension, "", "", size, hash]
+		  values = [@supplier, @dataset, filename, extension, "", "", size, hash, hash]
 		  yield(values)
 		end
 		break
@@ -320,7 +326,8 @@ module Embulk
 		eof = npart.to_s.rjust(3, "0")
 	      end
 	      content = Base64.encode64(data)
-	      values = [@supplier, @dataset, filename, extension, content, eof, size, hash]
+	      chunk_hash = @dataset_utils.getHashFromChunk(data)
+	      values = [@supplier, @dataset, filename, extension, content, eof, size, hash, chunk_hash]
 	      first = FALSE
 	      yield(values)
 	  end

ebulk-data/embulk-wendelin-dataset-tool/lib/embulk/input/wendelin.rb

@@ -242,7 +242,7 @@ module Embulk
       def initialize(task, schema, index, page_builder)
 	super
 	@data_set = task['data_set']
-	@chunk_size = task['chunk_size']
+	@chunk_size = DatasetUtils::CHUNK_SIZE
 	@data_set_directory = task['data_set_directory']
 	@wendelin = WendelinClient.new(task['erp5_url'], task['user'], task['password'])
 	@logger = LogManager.instance()
@@ -250,46 +250,52 @@ module Embulk
       end
 
       def run
-	data_stream = task['data_streams'][@index]
-	id = data_stream["id"]
-	reference = data_stream["reference"]
-	size = data_stream["size"]
-	hash = data_stream["hash"]
-	renamed = data_stream["status"] == DatasetUtils::STATUS_RENAMED
-	deleted = hash.to_s == DatasetUtils::DELETE
+	remote_file = task['data_streams'][@index]
+	reference = remote_file["reference"]
+	size = remote_file["full-size"]
+	large_hash = remote_file["large-hash"]
+	data_stream_chunk_list = remote_file["data-stream-list"]
+	renamed = remote_file["status"] == DatasetUtils::STATUS_RENAMED
+	deleted = large_hash.to_s == DatasetUtils::DELETE
 	begin
 	  if deleted
 	    entry = [reference, "", @data_set, DatasetUtils::DELETE, renamed]
 	    page_builder.add(entry)
 	  elsif renamed
-	    new_reference = data_stream["new_reference"]
+	    new_reference = remote_file["new_reference"]
 	    entry = [reference, new_reference, @data_set, TRUE, renamed]
   	    page_builder.add(entry)
 	  else
-	    @logger.info("Discarding local change on '#{data_stream["path"]}'", print=TRUE) if task['discard_changes']
+	    @logger.info("Discarding local change on '#{remote_file["path"]}'", print=TRUE) if task['discard_changes']
 	    @logger.info("Getting content from remote #{reference}", print=TRUE)
 	    @logger.info("Downloading...", print=TRUE)
-	    resume_split = @dataset_utils.splitOperationFileExist(reference) ? @dataset_utils.getLastSplitOperation(DatasetUtils::DOWNLOAD, reference, hash, @chunk_size) : 0
-	    n_chunk = resume_split == 0 ? 0 : resume_split+1
-	    split = n_chunk > 0
+	    resume_split = @dataset_utils.splitOperationFileExist(reference) ? @dataset_utils.getLastSplitOperation(DatasetUtils::DOWNLOAD, reference, large_hash, @chunk_size) : 0
+      n_chunk = resume_split == 0 ? 0 : resume_split+1
+      split = n_chunk > 0
 	    @logger.info("Resuming interrupted split download...", print=TRUE) if split
-	    @wendelin.eachDataStreamContentChunk(id, @chunk_size, n_chunk) do |chunk|
-	      content = chunk.nil? || chunk.empty? ? "" : Base64.encode64(chunk)
-	      begin_of_file = n_chunk == 0
-	      split = n_chunk > 0
-	      @dataset_utils.createSplitOperationControlFile(reference) if split
-	      entry = [reference, content, @data_set, begin_of_file, renamed]
-	      page_builder.add(entry)
-	      @dataset_utils.saveSplitOperation(DatasetUtils::DOWNLOAD, reference, n_chunk, hash, @chunk_size) if split
-	      n_chunk += 1
-	    end
+      data_stream_chunk_list.each_with_index do |data_stream_chunk, index|
+        #skip datastreams/chunks already downloaded
+        if n_chunk == index
+          content = ""
+          @wendelin.eachDataStreamContentChunk(data_stream_chunk["id"], @chunk_size, 0) do |chunk|
+            content = chunk.nil? || chunk.empty? ? "" : Base64.encode64(chunk)
+          end
+          begin_of_file = n_chunk == 0
+          split = n_chunk > 0
+          @dataset_utils.createSplitOperationControlFile(reference) if split
+          entry = [reference, content, @data_set, begin_of_file, renamed]
+          page_builder.add(entry)
+          @dataset_utils.saveSplitOperation(DatasetUtils::DOWNLOAD, reference, n_chunk, large_hash, @chunk_size) if split
+          n_chunk += 1
+        end
+      end
 	  end
 	  page_builder.finish
 	  @dataset_utils.deleteSplitOperationFile(reference) if split
 	rescue java.lang.OutOfMemoryError
 	  @logger.logOutOfMemoryError(reference)
 	  return_value = DatasetUtils::RUN_ABORTED
-	rescue Exception => e  
+	rescue Exception => e
 	  @logger.error(e.to_s, print=TRUE)
 	  @logger.error(e.backtrace)
 	  puts "[INFO] For more detailed information, please refer to the log file: " + @logger.getLogPath()
@@ -302,7 +308,9 @@ module Embulk
 	  if deleted
 	    @dataset_utils.deleteFromReport(reference, return_value)
 	  else
-	    @dataset_utils.addToReport(reference, return_value, size, hash, task['data_set'], new_reference)
+      file_path = renamed ? @dataset_utils.referenceToPath(new_reference, @data_set_directory, @data_set) : @dataset_utils.referenceToPath(reference, @data_set_directory, @data_set)
+      hash = @dataset_utils.getHash(file_path).to_s
+	    @dataset_utils.addToReport(reference, return_value, size, hash, large_hash, task['data_set'], new_reference)
 	  end
 	end
 	return {return_value => reference}