Ebulk changes

See merge request !1

ebulk

 #! /usr/bin/env bash
 
-DATA_LAKE_URL='https://wendelin.io/'
+DEFAULT_DATA_LAKE_URL='https://wendelin.io/'
 
-DOWN_URL="$DATA_LAKE_URL/"
-ING_POLICY="portal_ingestion_policies/default_embulk"
-ING_URL="$DATA_LAKE_URL$ING_POLICY"
+DOWN_URL="$DEFAULT_DATA_LAKE_URL"
+ING_POLICY="portal_ingestion_policies/default_ebulk"
+ING_URL="$DEFAULT_DATA_LAKE_URL$ING_POLICY"
 
-EBULK_VERSION="1.1.3"
+EBULK_VERSION="1.1.5"
 EMBULK_VERSION="0.9.7"
 EBULK_DATA_PATH=~/.ebulk
 EBULK_DATASET_FILE_NAME="/.ebulk_dataset"
@@ -49,7 +49,7 @@ ASK="A"
 if [ -f "$DATA_LAKE_URL_FILE" ]; then
   URL=$(cat "$DATA_LAKE_URL_FILE" 2>/dev/null)
   if [[ "$URL" != "" ]]; then
-    DOWN_URL="$URL/"
+    DOWN_URL="$URL"
     ING_URL="$URL$ING_POLICY"
   fi
 fi
@@ -221,13 +221,22 @@ function setDataLakeUrl {
       echo -e "[INFO] Please enter a valid url.${NC}"
       echo >&2; return 1
     fi
+    if [ "${URL: -1}" != "/" ] ; then
+      URL="$URL/"
+    fi
     echo "$URL" > "$DATA_LAKE_URL_FILE" 2>/dev/null
+    rm -f ${CREDENTIALS_FILE}
+    echo
+    echo "[INFO] Data-lake url set to '$URL'"
 }
 
 function defaultDataLakeUrl {
   echo "" > "$DATA_LAKE_URL_FILE" 2>/dev/null
-  DOWN_URL="$DATA_LAKE_URL/"
-  ING_URL="$DATA_LAKE_URL$ING_POLICY"
+  DOWN_URL="$DEFAULT_DATA_LAKE_URL"
+  ING_URL="$DEFAULT_DATA_LAKE_URL$ING_POLICY"
+  rm -f ${CREDENTIALS_FILE}
+  echo
+  echo "[INFO] Data-lake url set to default '$DEFAULT_DATA_LAKE_URL'"
 }
 
 function updateConfigFile {
@@ -305,9 +314,9 @@ function runProcess {
     if [ -z "$STATUS" ]; then
     	if [ ! -z "$CHUNK" ]; then
 	    if [ "$CHUNK" -eq "0" ]; then
-	    	echo "[INFO] Default chunk size: $DEFAULT_CHUNK_SIZE Mb."
+	    	echo "[INFO] Default chunk size: $DEFAULT_CHUNK_SIZE MB."
 	    else
-	    	echo "[INFO] Chunk size set in $CHUNK Mb."
+	    	echo "[INFO] Chunk size set in $CHUNK MB."
 	    fi
     	fi
 	if [ "$DATASET_DESCRIPTION" != "" ] ; then

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 		@logger.error("Your current dataset is outdated.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']}")
@@ -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

@@ -94,7 +94,6 @@ module Embulk
 	  @erp5_url = config.param('erp5_url', :string)
 	  @data_set = config.param('data_set', :string)
 	  @logger.info("Dataset name: #{@data_set}")
-	  @chunk_size = config.param('chunk_size', :float, default: 0) * DatasetUtils::MEGA
 	  @output_path = config.param("output_path", :string,  :default => nil)
 	  if not File.directory?(@output_path)
 	    @logger.error("Output directory not found.", print=TRUE)
@@ -103,14 +102,14 @@ module Embulk
 	  task = {
 		 'erp5_url' => @erp5_url,
 		 'data_set' => @data_set,
-		 'chunk_size' => @chunk_size,
+		 'chunk_size' => DatasetUtils::CHUNK_SIZE + 10,
 		 'output_path' => @output_path,
 		 'tool_dir' => @tool_dir
 	       }
 	  if task['chunk_size'] == 0
 	    task['chunk_size'] = DatasetUtils::CHUNK_SIZE
 	  end
-	  @logger.info("Chunk size set in #{task['chunk_size']/DatasetUtils::MEGA}MB")
+	  @logger.info("Download chunk size relies on server file chunks.")
 	  @dataset_utils = DatasetUtils.new("")
 	  task['data_set_directory'] = @dataset_utils.appendSlashTo(@output_path)
 	  @data_set_directory = task['data_set_directory']
@@ -242,7 +241,6 @@ module Embulk
       def initialize(task, schema, index, page_builder)
 	super
 	@data_set = task['data_set']
-	@chunk_size = task['chunk_size']
 	@data_set_directory = task['data_set_directory']
 	@wendelin = WendelinClient.new(task['erp5_url'], task['user'], task['password'])
 	@logger = LogManager.instance()
@@ -250,46 +248,61 @@ 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']
+      chunk_size = data_stream_chunk_list[0]["size"] #first chunk size
+	    @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
-	    @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
+	    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
+      if split
+        @logger.info("Resuming interrupted split download...", print=TRUE)
+      else
+        if data_stream_chunk_list.length > 1
+          @logger.info("Downloading large file split in chunks...", print=TRUE)
+        else
+          @logger.info("Downloading...", print=TRUE)
+        end
+      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 + 10, 0, data_stream_chunk_list.length > 1) 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
+      @logger.info("Done", print=TRUE) if data_stream_chunk_list.length > 1
 	  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 +315,9 @@ 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}

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

@@ -60,7 +60,7 @@ module Embulk
 	      File.open(file_path, write_mode) { |file| file.write(data_chunk) }
 	    end
 	  end
-	rescue Exception => e  
+	rescue Exception => e
 	  @logger.error("An error occurred while procesing file.", print=TRUE)
 	  @logger.error(e.backtrace)
 	  raise e

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

@@ -34,7 +34,7 @@ class WendelinClient
 
   def exists(reference)
     checkReferenceChars(reference)
-    uri = URI(URI.escape("#{@erp5_url}/ingestionReferenceExists?reference=#{reference}"))
+    uri = URI(URI.escape("#{@erp5_url}/ERP5Site_checkIngestionReferenceExists?reference=#{reference}"))
     begin
       response = handleRequest(uri)
     rescue Exception => e
@@ -139,14 +139,14 @@ class WendelinClient
       return {"success"=>TRUE, "message"=>"success"}
   end
 
-  def eachDataStreamContentChunk(id, chunk_size, n_chunk=0)
+  def eachDataStreamContentChunk(id, chunk_size, n_chunk=0, split_operation=FALSE)
     n_part = n_chunk
     done = FALSE
     first = TRUE
     while not done
       start_offset = n_part*chunk_size
       end_offset = n_part*chunk_size+chunk_size
-      uri = URI(URI.escape("#{@erp5_url}getDataStreamChunk?id=#{id}&start_offset=#{start_offset}&end_offset=#{end_offset}"))
+      uri = URI(URI.escape("#{@erp5_url}ERP5Site_getDataStreamChunk?id=#{id}&start_offset=#{start_offset}&end_offset=#{end_offset}"))
       success = FALSE
       n_retry = 0
       while ! success && n_retry < 10
@@ -158,7 +158,11 @@ class WendelinClient
 		if first
 		  yield chunk
 		end
-	        @logger.info("Done", print=TRUE)
+	        if split_operation
+	          @logger.info("File chunk downloaded", print=TRUE)
+	        else
+	          @logger.info("Done", print=TRUE)
+	        end
 	        done = TRUE
 	      else
 		first = FALSE
@@ -181,7 +185,7 @@ class WendelinClient
   end
 
   def getDataStreams(data_set_reference)
-    uri = URI(URI.escape("#{@erp5_url}getDataStreamList?data_set_reference=#{data_set_reference}"))
+    uri = URI(URI.escape("#{@erp5_url}ERP5Site_getDataStreamList?data_set_reference=#{data_set_reference}"))
     response = handleRequest(uri)
     if response["success"] == FALSE
       @logger.abortExecution()
@@ -235,6 +239,8 @@ class WendelinClient
         return {"success"=>TRUE, "message"=>res.body}
       else
         @logger.error("HTTP FAIL -  code: #{res.code}", print=TRUE)
+        @logger.error("During request to " + uri.hostname.to_s, print=TRUE)
+        @logger.error(uri.to_s)
         if res.code == '500' or res.code == '502' or res.code == '503'
           @logger.error(HTTP_MESSAGE_5XX, print=TRUE)
         elsif res.code == '401'

ebulk-data/help.md

@@ -16,6 +16,7 @@ commands:
    -h,  --help           Tool help
    -r,  --readme         Opens README file
    -e,  --examples       Shows some tool usage examples
+   -v,  --version        Ebulk tool version
    store-credentials	 Stores user and password for automatic authentication
    set-data-lake-url	 Sets the data lake url where to ingest/download
    default-data-lake-url Sets the data lake url to default