oai_schedules 0.7.0 → 0.8.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 93b5d732552fa18fc37680009dd21bf9509555c5cfeec8c26b2b20a37967f606
-  data.tar.gz: cb282dfcd7e8866faa698e5f310ba07223e93d965395734a26665e531cca3ce0
+  metadata.gz: 86d2572fbf72c1d75ab9b83efcbb739726962127953458f438cf5ac20278997d
+  data.tar.gz: b4fe62d6a8280e9bd55c2337d10615768808d0d5ccbeef469a8ad4a7a55b0b05
 SHA512:
-  metadata.gz: 54a2044fe8446bf0951415302b68d06dccb3c36fc2628be888681418d1a4cecfcfd2723adbce8dd61656a3081087516e40c28e40acaa9037fdbeda88797f9d28
-  data.tar.gz: 553f101400c60d6f1134f8ea97c075183a595059737605ab49ede6ef8620a5e0b15d47b2f82ff4253036c485489ddebc0059074f3d87bc3c2fc523d98e067600
+  metadata.gz: 1294b5a37cf3d9ef098c5d1484d1853e5c5990148348b735c7a22809cbab963c1d22e1fae8b3ab525107a21368483ac8eb600f954293dde41a6039db4fed6097
+  data.tar.gz: bc8a60e47cd6f1778483556e644362e976b97456b5c30ae6c320b526b55958e2852011c002c3268afd15c003f835a03b847d0a9173f4590e418ccfece198f447

data/CHANGELOG.md

@@ -1,3 +1,8 @@
+## [0.8.0] - 2025-07-10
+
+- Complete harvesting is now within a single thread tick
+- Added more info in the schedule state
+
 ## [0.7.0] - 2025-07-09
 
 - Partial harvest policy changed: avoid using resumption tokens

data/lib/oai_schedules/manager.rb

@@ -231,11 +231,14 @@ module OAISchedules
       state["resumption_token"] = nil
       state["expiration_date_resumption_token"] = nil
       state["datetime_now"] = get_datetime_now
+      state["datetime_next_harvesting"] = nil
       state["count_success"] = 0
       state["count_fails"] = 0
       state["done"] = false
       state["count_harvested_records"] = 0
       state["latest_harvested_records_datestamp"] = nil
+      state["harvesting"] = false
+      state["error"] = nil
     end
 
 
@@ -259,8 +262,7 @@ module OAISchedules
       task = @schedules[name][:task]
       interval_s_safe = 60
       begin
-        duration = ISO8601::Duration.new(@schedules[name][:content]["interval"])
-        interval_s = duration.to_seconds
+        interval_s = interval_iso8601_to_seconds(@schedules[name][:content]["interval"])
         @logger.info("#{name}: task interval (s): #{interval_s}")
         th_interval_s = 0.1   # protects from negative, 0 or small time intervals
         if interval_s < th_interval_s
@@ -338,7 +340,9 @@ module OAISchedules
 
 
     def logic(name, content, state_machine, state)
+      state["datetime_next_harvesting"] = get_datetime_next_schedule_tick_from_now(content["interval"])
       loop do
+        # sleep(3)
         @logger.info("#{name}: handling state: #{state_machine.state}")
         case state_machine.state
         when StateHarvesting::NOT_IDENTIFIED
@@ -353,7 +357,8 @@ module OAISchedules
         when StateHarvesting::IDLE
           state_machine.add_event(EventHarvesting::REQUEST_HARVEST)
         when StateHarvesting::HARVESTING
-          use_resumption_token = false   # don't change
+          use_resumption_token = false # don't change
+          state["harvesting"] = true
           format = content["format"] || ""
           fmt_dt = content.dig("repository", "granularity")
           from = convert_datetime(content["from"], fmt_dt) || ""
@@ -400,7 +405,6 @@ module OAISchedules
                   state["expiration_date_resumption_token"] = data_token["_expirationDate"]
                 end
               end
-              state["datetime_now"] = get_datetime_now
               n_records = data["record"].size
               state["count_harvested_records"] += n_records
               if n_records > 0
@@ -416,6 +420,7 @@ module OAISchedules
               error = e
             end
             if error.nil?
+              state["error"] = nil
               if use_resumption_token
                 if state["resumption_token"].nil?
                   state_machine.add_event(EventHarvesting::DONE_HARVEST)
@@ -424,11 +429,19 @@ module OAISchedules
                   to_pause = false
                 end
               else
-                state_machine.add_event(EventHarvesting::DONE_HARVEST)
-                to_pause = true
+                to_pause = false
               end
+            else
+              state["error"] = {
+                "message" => error.message,
+                "backtrace" => error.backtrace
+              }
+              state_machine.add_event(EventHarvesting::DONE_HARVEST)
+              to_pause = true
             end
             state["done"] = done
+            state["harvesting"] = !to_pause
+            state["datetime_now"] = get_datetime_now
           end
           path_file_state = get_path_state_file_from_schedule_name(name)
           @logger.info("#{name}: writing to state file #{path_file_state}")
@@ -570,6 +583,17 @@ module OAISchedules
       DateTime.now.strftime('%FT%TZ')
     end
 
+    def interval_iso8601_to_seconds(str_interval)
+      ISO8601::Duration.new(str_interval).to_seconds
+    end
+
+    def get_datetime_next_schedule_tick_from_now(str_interval)
+      interval_s = interval_iso8601_to_seconds(str_interval)
+      dt = DateTime.now
+      dt += 1.0*interval_s/(24*60*60)
+      dt.strftime('%FT%TZ')
+    end
+
   end
 
 end

data/lib/oai_schedules/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module OaiSchedules
-  VERSION = "0.7.0"
+  VERSION = "0.8.0"
 end

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: oai_schedules
 version: !ruby/object:Gem::Version
-  version: 0.7.0
+  version: 0.8.0
 platform: ruby
 authors:
 - Davide Monari
 bindir: exe
 cert_chain: []
-date: 2025-07-09 00:00:00.000000000 Z
+date: 2025-07-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: concurrent-ruby