[IMP] start extraction of chunk processing \o/

Author: sebastienbeau

chunk_processing/__init__.py

@@ -0,0 +1,2 @@
+from . import models
+from . import components

chunk_processing/__manifest__.py

@@ -0,0 +1,28 @@
+# Copyright 2021 Akretion (https://www.akretion.com).
+# @author Sébastien BEAU <sebastien.beau@akretion.com>
+# License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+
+
+{
+    "name": "Chunk Processing",
+    "summary": "Base module for processing chunk",
+    "version": "14.0.1.0.0",
+    "category": "Uncategorized",
+    "website": "https://github.com/shopinvader/pattern-import-export",
+    "author": " Akretion",
+    "license": "AGPL-3",
+    "application": False,
+    "installable": True,
+    "external_dependencies": {
+        "python": [],
+        "bin": [],
+    },
+    "depends": [
+        "queue_job",
+        "component",
+    ],
+    "data": [
+        "views/chunk_item_view.xml",
+    ],
+    "demo": [],
+}

chunk_processing/components/__init__.py

@@ -0,0 +1,3 @@
+from . import processor
+from . import splitter
+from . import splitter_json

chunk_processing/components/processor.py

@@ -0,0 +1,14 @@
+# Copyright 2021 Akretion (https://www.akretion.com).
+# @author Sébastien BEAU <sebastien.beau@akretion.com>
+# License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+
+
+from odoo.addons.component.core import AbstractComponent
+
+
+class ChunkProcessor(AbstractComponent):
+    _name = "chunk.processor"
+    _collection = "chunk.item"
+
+    def run(self):
+        raise NotImplementedError

chunk_processing/components/splitter.py

@@ -0,0 +1,50 @@
+# Copyright 2021 Akretion (https://www.akretion.com).
+# @author Sébastien BEAU <sebastien.beau@akretion.com>
+# License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+
+from odoo.addons.component.core import AbstractComponent
+
+
+class ChunkSplitter(AbstractComponent):
+    _name = "chunk.splitter"
+    _collection = "chunk.group"
+
+    def _parse_data(self, data):
+        raise NotImplementedError
+
+    def _prepare_chunk(self, start_idx, stop_idx, data):
+        return {
+            "start_idx": start_idx,
+            "stop_idx": stop_idx,
+            "data": data,
+            "nbr_item": len(data),
+            "state": "pending",
+            "group_id": self.collection.id,
+        }
+
+    def _should_create_chunk(self, items, next_item):
+        """Customise this code if you want to add some additionnal
+        item after reaching the limit"""
+        return len(items) > self.collection.chunk_size
+
+    def _create_chunk(self, start_idx, stop_idx, data):
+        vals = self._prepare_chunk(start_idx, stop_idx, data)
+        chunk = self.env["chunk.item"].create(vals)
+        # we enqueue the chunk in case of multi process of if it's the first chunk
+        if self.collection.process_multi or len(self.collection.item_ids) == 1:
+            chunk.with_delay(priority=self.collection.job_priority).run()
+        return chunk
+
+    def run(self, data):
+        items = []
+        start_idx = 1
+        previous_idx = None
+        for idx, item in self._parse_data(data):
+            if self._should_create_chunk(items, item):
+                self._create_chunk(start_idx, previous_idx, items)
+                items = []
+                start_idx = idx
+            items.append((idx, item))
+            previous_idx = idx
+        if items:
+            self._create_chunk(start_idx, idx, items)

chunk_processing/components/splitter_json.py

@@ -0,0 +1,18 @@
+# Copyright 2021 Akretion (https://www.akretion.com).
+# @author Sébastien BEAU <sebastien.beau@akretion.com>
+# License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+
+import json
+
+from odoo.addons.component.core import Component
+
+
+class ChunkSplitterJson(Component):
+    _inherit = "chunk.splitter"
+    _name = "chunk.splitter.json"
+    _usage = "json"
+
+    def _parse_data(self, data):
+        items = json.loads(data.decode("utf-8"))
+        for idx, item in enumerate(items):
+            yield idx + 1, item

chunk_processing/models/__init__.py

@@ -0,0 +1,2 @@
+from . import chunk_item
+from . import chunk_group

chunk_processing/models/chunk_group.py

@@ -0,0 +1,69 @@
+# Copyright 2021 Akretion (https://www.akretion.com).
+# @author Sébastien BEAU <sebastien.beau@akretion.com>
+# License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+
+
+from odoo import _, api, fields, models
+
+
+class ChunkGroup(models.Model):
+    _inherit = "collection.base"
+    _name = "chunk.group"
+
+    item_ids = fields.One2many("chunk.item", "group_id", "Item")
+    process_multi = fields.Boolean()
+    job_priority = fields.Integer(default=20)
+    chunk_size = fields.Integer(default=500, help="Define the size of the chunk")
+    progress = fields.Float(compute="_compute_stat")
+    date_done = fields.Datetime()
+    data_format = fields.Selection(
+        [
+            ("json", "Json"),
+            ("xml", "XML"),
+        ]
+    )
+    state = fields.Selection(
+        [("pending", "Pending"), ("failed", "Failed"), ("done", "Done")],
+        default="pending",
+    )
+    info = fields.Char()
+    nbr_error = fields.Integer(compute="_compute_stat")
+    nbr_success = fields.Integer(compute="_compute_stat")
+    apply_on_model = fields.Char()
+    usage = fields.Char()
+
+    @api.depends("item_ids.nbr_error", "item_ids.nbr_success")
+    def _compute_stat(self):
+        for record in self:
+            record.nbr_error = sum(record.mapped("item_ids.nbr_error"))
+            record.nbr_success = sum(record.mapped("item_ids.nbr_success"))
+            todo = sum(record.mapped("item_ids.nbr_item"))
+            if todo:
+                record.progress = (record.nbr_error + record.nbr_success) * 100.0 / todo
+            else:
+                record.progress = 0
+
+    def _get_data(self):
+        raise NotImplementedError
+
+    def split_in_chunk(self):
+        """Split Group into Chunk"""
+        # purge chunk in case of retring a job
+        self.item_ids.unlink()
+        try:
+            data = self._get_data()
+            with self.work_on(self._name) as work:
+                splitter = work.component(usage=self.data_format)
+                splitter.run(data)
+        except Exception as e:
+            self.state = "failed"
+            self.info = _("Failed to create the chunk: %s") % e
+        return True
+
+    def set_done(self):
+        for record in self:
+            if record.nbr_error:
+                record.state = "failed"
+            else:
+                record.state = "done"
+            record.date_done = fields.Datetime.now()

…rn_import_export/models/pattern_chunk.py chunk_processing/models/chunk_item.pypattern_import_export/models/pattern_chunk.py renamed to chunk_processing/models/chunk_item.py pattern_import_export/models/pattern_chunk.py renamed to chunk_processing/models/chunk_item.py

@@ -5,14 +5,15 @@
 from odoo import fields, models
 
 
-class PatternChunk(models.Model):
-    _name = "pattern.chunk"
-    _description = "Pattern Chunk"
+class ChunkItem(models.Model):
+    _inherit = "collection.base"
+    _name = "chunk.item"
+    _description = "Chunk Item"
     _order = "start_idx"
     _rec_name = "start_idx"
 
-    pattern_file_id = fields.Many2one(
-        "pattern.file", "Pattern File", required=True, ondelete="cascade"
+    group_id = fields.Many2one(
+        "chunk.group", "Chunk Group", required=True, ondelete="cascade"
     )
     start_idx = fields.Integer()
     stop_idx = fields.Integer()
@@ -32,33 +33,31 @@ class PatternChunk(models.Model):
         ]
     )
 
-    def run_import(self):
-        model = self.pattern_file_id.pattern_config_id.model_id.model
-        res = (
-            self.with_context(pattern_config={"model": model, "record_ids": []})
-            .env[model]
-            .load([], self.data)
-        )
-        self.write(self._prepare_chunk_result(res))
-        config = self.pattern_file_id.pattern_config_id
-        priority = config.job_priority
-        if not config.process_multi:
+    def manual_run(self):
+        """ Run the import without try/except, easier for debug """
+        return self._run()
+
+    def _run(self):
+        with self.work_on(self.group_id.apply_on_model) as work:
+            processor = work.component(usage=self.group_id.usage)
+            processor.run()
+        if not self.group_id.process_multi:
             next_chunk = self.get_next_chunk()
             if next_chunk:
-                next_chunk.with_delay(priority=priority).run()
+                next_chunk.with_delay(priority=self.group_id.job_priority).run()
             else:
                 self.with_delay(priority=5).check_last()
         else:
             self.with_delay(priority=5).check_last()
 
     def run(self):
-        """Process Import of Pattern Chunk"""
+        """Process Chunk Item in a savepoint"""
         cr = self.env.cr
         try:
             self.state = "started"
             cr.commit()  # pylint: disable=invalid-commit
             with cr.savepoint():
-                self.run_import()
+                self._run()
         except Exception as e:
             self.write(
                 {
@@ -70,6 +69,7 @@ def run(self):
             self.with_delay().check_last()
         return "OK"
 
+    # TODO move this in pattern-import
     def _prepare_chunk_result(self, res):
         # TODO rework this part and add specific test case
         nbr_error = len(res["messages"])
@@ -98,23 +98,19 @@ def _prepare_chunk_result(self, res):
         }
 
     def get_next_chunk(self):
-        return self.search(
-            [
-                ("pattern_file_id", "=", self.pattern_file_id.id),
-                ("state", "=", "pending"),
-            ],
-            limit=1,
+        return fields.first(
+            self.group_id.item_ids.filtered(lambda s: s.state == "pending")
         )
 
     def is_last_job(self):
-        return not self.pattern_file_id.chunk_ids.filtered(
+        return not self.group_id.item_ids.filtered(
             lambda s: s.state in ("pending", "started")
         )
 
     def check_last(self):
         """Check if all chunk have been processed"""
         if self.is_last_job():
-            self.pattern_file_id.set_import_done()
-            return "Pattern file is done"
+            self.group_id.set_done()
+            return "Chunk group is done"
         else:
             return "There is still some running chunk"

…rn_import_export/views/pattern_chunk.xml …unk_processing/views/chunk_item_view.xmlpattern_import_export/views/pattern_chunk.xml renamed to chunk_processing/views/chunk_item_view.xml pattern_import_export/views/pattern_chunk.xml renamed to chunk_processing/views/chunk_item_view.xml

@@ -1,8 +1,8 @@
 <?xml version="1.0" encoding="UTF-8" ?>
 <odoo>
 
-<record id="pattern_chunk_view_tree" model="ir.ui.view">
-    <field name="model">pattern.chunk</field>
+<record id="chunk_item_view_tree" model="ir.ui.view">
+    <field name="model">chunk.item</field>
     <field name="arch" type="xml">
         <tree string="Chunk">
             <field name="start_idx" />
@@ -14,12 +14,12 @@
     </field>
 </record>
 
-<record id="pattern_chunk_view_form" model="ir.ui.view">
-    <field name="model">pattern.chunk</field>
+<record id="chunk_item_view_form" model="ir.ui.view">
+    <field name="model">chunk.item</field>
     <field name="arch" type="xml">
         <form string="Chunk">
             <header>
-                <button name="run_import" type="object" string="Force manual run" />
+                <button name="manual_run" type="object" string="Force manual run" />
             </header>
             <sheet>
                 <group>