Introducing WorkUnit model names (#98)

Now `WorkUnit` fields `name: str` and `measurement: str` moves to `model_name: dict[str, Any]`, also `name` becomes `granular_name`.

This introduces universal format for works unique identification.

Author: StannisMod

examples/simple_generation.ipynb

@@ -99,6 +99,41 @@
     "                                     uniq_resources=100)"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "{1: 'hello'}\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": "dict"
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "d = {1: 'hello'}\n",
+    "print(d)\n",
+    "\n",
+    "v = eval(str(d))\n",
+    "type(v)"
+   ],
+   "metadata": {
+    "collapsed": false,
+    "ExecuteTime": {
+     "end_time": "2025-09-17T08:15:05.619946Z",
+     "start_time": "2025-09-17T08:15:05.611943300Z"
+    }
+   }
+  },
   {
    "cell_type": "markdown",
    "metadata": {

examples/simple_synthetic_graph_scheduling.py

@@ -31,7 +31,7 @@
 
 # Get information about created WorkGraph's attributes
 works_count = len(wg.nodes)
-work_names_count = len(set(n.work_unit.name for n in wg.nodes))
+work_names_count = len(set(n.work_unit.model_name for n in wg.nodes))
 res_kind_count = len(set(req.kind for req in chain(*[n.work_unit.worker_reqs for n in wg.nodes])))
 print(works_count, work_names_count, res_kind_count)
 

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "sampo"
-version = "0.1.1.353"
+version = "0.1.2"
 description = "Open-source framework for adaptive manufacturing processes scheduling"
 authors = ["iAirLab <[email protected]>"]
 license = "BSD-3-Clause"

sampo/generator/pipeline/cluster.py

@@ -17,15 +17,18 @@ def _add_addition_work(probability: float, rand: Random | None = None) -> bool:
 def _get_roads(parents: list[GraphNode], cluster_name: str, dist: float,
                rand: Random | None = None) -> tuple[dict[str, GraphNode], int]:
     road_nodes = dict()
-    min_r = WorkUnit(uuid_str(rand), 'minimal road',
-                     scale_reqs(wr.MIN_ROAD, dist), group=f'{cluster_name}:road', volume=dist, volume_type='km')
+    min_r = WorkUnit(uuid_str(rand), {'granular_name': 'minimal road',
+                                      'measurement': 'km'},
+                     scale_reqs(wr.MIN_ROAD, dist), group=f'{cluster_name}:road', volume=dist)
     road_nodes['min'] = GraphNode(min_r, parents)
-    temp_r = WorkUnit(uuid_str(rand), 'temporary road',
-                      scale_reqs(wr.TEMP_ROAD, dist), group=f'{cluster_name}:road', volume=dist, volume_type='km')
+    temp_r = WorkUnit(uuid_str(rand), {'granular_name': 'temporary road',
+                                       'measurement': 'km'},
+                      scale_reqs(wr.TEMP_ROAD, dist), group=f'{cluster_name}:road', volume=dist)
     road_nodes['temp'] = GraphNode(temp_r, [(road_nodes['min'], wr.ATOMIC_ROAD_LEN, EdgeType.LagFinishStart)])
 
-    final_r = WorkUnit(uuid_str(rand), 'final road', scale_reqs(wr.FINAL_ROAD, dist), group=f'{cluster_name}:road',
-                       volume=dist, volume_type='km')
+    final_r = WorkUnit(uuid_str(rand), {'granular_name': 'final road',
+                                        'measurement': 'km'},
+                       scale_reqs(wr.FINAL_ROAD, dist), group=f'{cluster_name}:road', volume=dist)
     road_nodes['final'] = GraphNode(final_r, [(road_nodes['temp'], wr.ATOMIC_ROAD_LEN, EdgeType.LagFinishStart)])
     return road_nodes, len(road_nodes)
 
@@ -43,21 +46,27 @@ def _get_engineering_preparation(parents: list[GraphNode], cluster_name: str, bo
 def _get_power_lines(parents: list[GraphNode], cluster_name: str, dist_line: float,
                      dist_high_line: float | None = None, rand: Random | None = None) -> tuple[list[GraphNode], int]:
     worker_req = wr.scale_reqs(wr.POWER_LINE, dist_line)
-    power_line_1 = WorkUnit(uuid_str(rand), 'power line', worker_req,
+    power_line_1 = WorkUnit(uuid_str(rand),
+                            {'granular_name': 'power line', 'measurement': 'km'},
+                            worker_req,
                             group=f'{cluster_name}:electricity',
-                            volume=dist_line, volume_type='km')
-    power_line_2 = WorkUnit(uuid_str(rand), 'power line', worker_req,
+                            volume=dist_line)
+    power_line_2 = WorkUnit(uuid_str(rand),
+                            {'granular_name': 'power line', 'measurement': 'km'},
+                            worker_req,
                             group=f'{cluster_name}:electricity',
-                            volume=dist_line, volume_type='km')
+                            volume=dist_line)
 
     power_lines = [
         GraphNode(power_line_1, parents),
         GraphNode(power_line_2, parents),
     ]
     if dist_high_line is not None:
         worker_req_high = wr.scale_reqs(wr.POWER_LINE, dist_high_line)
-        high_power_line = WorkUnit(uuid_str(rand), 'high power line', worker_req_high,
-                                   group=f'{cluster_name}:electricity', volume=dist_high_line, volume_type='km')
+        high_power_line = WorkUnit(uuid_str(rand),
+                                   {'granular_name': 'high power line', 'measurement': 'km'},
+                                   worker_req_high,
+                                   group=f'{cluster_name}:electricity', volume=dist_high_line)
         power_lines.append(GraphNode(high_power_line, parents))
 
     return power_lines, len(power_lines)
@@ -66,23 +75,32 @@ def _get_power_lines(parents: list[GraphNode], cluster_name: str, dist_line: flo
 def _get_pipe_lines(parents: list[GraphNode], cluster_name: str, pipe_dists: list[float],
                     rand: Random | None = None) -> tuple[list[GraphNode], int]:
     worker_req_pipe = wr.scale_reqs(wr.PIPE_LINE, pipe_dists[0])
-    first_pipe = WorkUnit(uuid_str(rand), 'pipe', worker_req_pipe, group=f'{cluster_name}:oil_gas_long_pipes',
-                          volume=pipe_dists[0], volume_type='km')
+    first_pipe = WorkUnit(uuid_str(rand),
+                          {'granular_name': 'pipe', 'measurement': 'km'},
+                          worker_req_pipe, group=f'{cluster_name}:oil_gas_long_pipes',
+                          volume=pipe_dists[0])
 
     graph_nodes = [GraphNode(first_pipe, parents)]
     for i in range(1, len(pipe_dists)):
-        node_work = WorkUnit(uuid_str(rand), 'node', wr.PIPE_NODE,
+        node_work = WorkUnit(uuid_str(rand),
+                             {'granular_name': 'node'},
+                             wr.PIPE_NODE,
                              group=f'{cluster_name}:oil_gas_long_pipes')
         graph_nodes.append(GraphNode(node_work, parents))
         worker_req_pipe = wr.scale_reqs(wr.PIPE_LINE, pipe_dists[i])
-        pipe_work = WorkUnit(uuid_str(rand), 'pipe', worker_req_pipe,
+        pipe_work = WorkUnit(uuid_str(rand),
+                             {'granular_name': 'pipe', 'measurement': 'km'},
+                             worker_req_pipe,
                              group=f'{cluster_name}:oil_gas_long_pipes',
-                             volume=pipe_dists[i], volume_type='km')
+                             volume=pipe_dists[i])
         graph_nodes.append(GraphNode(pipe_work, parents))
 
     worker_req_loop = wr.scale_reqs(wr.PIPE_LINE, pipe_dists[0])
-    looping = WorkUnit(uuid_str(rand), 'looping', worker_req_loop, group=f'{cluster_name}:oil_gas_long_pipes',
-                       volume=pipe_dists[0], volume_type='km')
+    looping = WorkUnit(uuid_str(rand),
+                       {'granular_name': 'looping', 'measurement': 'km'},
+                       worker_req_loop,
+                       group=f'{cluster_name}:oil_gas_long_pipes',
+                       volume=pipe_dists[0])
     graph_nodes.append(GraphNode(looping, graph_nodes[0:1]))
     return graph_nodes, len(graph_nodes)
 
@@ -147,18 +165,23 @@ def _get_boreholes_equipment_general(parents: list[GraphNode], cluster_name: str
         dist = gen_c.DIST_BETWEEN_BOREHOLES.rand_float(rand)
         dists_sum += dist
         worker_req_pipe = scale_reqs(wr.POWER_NETWORK, dist)
-        pipe_net_work = WorkUnit(uuid_str(rand), 'elem of pipe_network', worker_req_pipe,
-                                 group=f'{cluster_name}:oil_gas_pipe_net', volume=dist, volume_type='km')
+        pipe_net_work = WorkUnit(uuid_str(rand),
+                                 {'granular_name': 'elem of pipe_network', 'measurement': 'km'},
+                                 worker_req_pipe,
+                                 group=f'{cluster_name}:oil_gas_pipe_net',
+                                 volume=dist)
         nodes.append(GraphNode(pipe_net_work, parents))
 
     worker_req_power = scale_reqs(wr.POWER_NETWORK, dists_sum)
-    power_net_work = WorkUnit(uuid_str(rand), 'power network', worker_req_power,
+    power_net_work = WorkUnit(uuid_str(rand), {'granular_name': 'power network', 'measurement': 'km'},
+                              worker_req_power,
                               group=f'{cluster_name}:electricity',
-                              volume=dists_sum, volume_type='km')
+                              volume=dists_sum)
     nodes.append(GraphNode(power_net_work, parents))
 
     for i in range(masts_count):
-        light_mast_work = WorkUnit(uuid_str(rand), 'mast', wr.LIGHT_MAST,
+        light_mast_work = WorkUnit(uuid_str(rand), {'granular_name': 'mast', 'measurement': 'km'},
+                                   wr.LIGHT_MAST,
                                    group=f'{cluster_name}:light_masts')
         nodes.append(GraphNode(light_mast_work, parents))
     return nodes, len(nodes)

sampo/scheduler/multi_agency/block_generator.py

@@ -93,7 +93,7 @@ def generate_wg(mode, i):
             bg.add_edge(global_start, node)
             bg.add_edge(node, global_end)
 
-    logger(f'{graph_type.name} ' + ' '.join([str(mode.name) for i, mode in enumerate(modes)
+    logger(f'{graph_type.name} ' + ' '.join([str(mode.model_name) for i, mode in enumerate(modes)
                                              if nodes[i].vertex_count != EMPTY_GRAPH_VERTEX_COUNT]))
     return bg
 

sampo/scheduler/utils/local_optimization.py

@@ -247,16 +247,16 @@ def optimize(self, scheduled_works: dict[GraphNode, ScheduledWork], node_order:
                 satisfy = True
 
                 for candidate_worker in candidate_schedule.workers:
-                    my_worker = my_workers.get(candidate_worker.name, None)
+                    my_worker = my_workers.get(candidate_worker.model_name, None)
                     if my_worker is None:  # these two works are not compete for this worker
                         continue
 
-                    need_me = my_workers[candidate_worker.name].count
+                    need_me = my_workers[candidate_worker.model_name].count
                     need_candidate = candidate_worker.count
 
                     total = need_me + need_candidate
-                    my_req = my_schedule_reqs[candidate_worker.name]
-                    candidate_req = candidate_schedule_reqs[candidate_worker.name]
+                    my_req = my_schedule_reqs[candidate_worker.model_name]
+                    candidate_req = candidate_schedule_reqs[candidate_worker.model_name]
                     needed_min = my_req.min_count + candidate_req.min_count
 
                     if needed_min > total:  # these two works can't run in parallel
@@ -273,17 +273,17 @@ def optimize(self, scheduled_works: dict[GraphNode, ScheduledWork], node_order:
                     my_worker_count += add_me
                     candidate_worker_count += add_candidate
 
-                    new_my_workers[candidate_worker.name] = my_worker_count
-                    new_candidate_workers[candidate_worker.name] = candidate_worker_count
+                    new_my_workers[candidate_worker.model_name] = my_worker_count
+                    new_candidate_workers[candidate_worker.model_name] = candidate_worker_count
 
                 if satisfy:  # replacement found, apply changes and leave candidates bruteforce
                     print(f'Found! {candidate.work_unit.name} {node.work_unit.name}')
                     for worker in my_schedule.workers:
-                        worker_count = new_my_workers.get(worker.name, None)
+                        worker_count = new_my_workers.get(worker.model_name, None)
                         if worker_count is not None:
                             worker.count = worker_count
                     for worker in candidate_schedule.workers:
-                        worker_count = new_candidate_workers.get(worker.name, None)
+                        worker_count = new_candidate_workers.get(worker.model_name, None)
                         if worker_count is not None:
                             worker.count = worker_count
                     # candidate_schedule.start_time = my_schedule.start_time

sampo/schemas/schedule.py

@@ -1,28 +1,34 @@
 from copy import deepcopy
 from datetime import datetime
 from functools import lru_cache
-from typing import Iterable, Union
+from operator import itemgetter
+from typing import Iterable, Union, Any
 
 from pandas import DataFrame
 
 from sampo.schemas.graph import WorkGraph, GraphNode
 from sampo.schemas.scheduled_work import ScheduledWork
 from sampo.schemas.serializable import JSONSerializable, T
 from sampo.schemas.time import Time
+from sampo.utilities.collections_util import first
 from sampo.utilities.schedule import fix_split_tasks, offset_schedule
 
 ResourceSchedule = dict[str, list[tuple[Time, Time]]]
 ScheduleWorkDict = dict[str, ScheduledWork]
 
 
+def _get_granular_name_columns(sworks: Iterable[ScheduledWork] | None = None):
+    return list(sorted(first(sworks).model_name.keys()))
+
+
 # TODO: Rebase object onto ScheduleWorkDict and ordered ScheduledWork list
 class Schedule(JSONSerializable['Schedule']):
     """
     Represents work schedule. Is a wrapper around DataFrame with specific structure.
     """
 
-    _data_columns: list[str] = ['idx', 'task_id', 'task_name', 'task_name_mapped', 'contractor', 'cost',
-                                'volume', 'measurement', 'start',
+    _data_columns: list[str] = ['idx', 'task_id', 'task_name', 'contractor', 'cost',
+                                'volume', 'start',
                                 'finish', 'duration', 'workers']
     _scheduled_work_column: str = 'scheduled_work_object'
 
@@ -47,7 +53,7 @@ def pure_schedule_df(self) -> DataFrame:
         return self._schedule[~self._schedule.apply(
             lambda row: row[self._scheduled_work_column].is_service_unit,
             axis=1
-        )][self._data_columns]
+        )][self._data_columns + _get_granular_name_columns(self._schedule[self._scheduled_work_column])]
 
     @property
     def works(self) -> Iterable[ScheduledWork]:
@@ -118,7 +124,7 @@ def unite_stages(self) -> 'Schedule':
         def f(row):
             swork: ScheduledWork = deepcopy(row[self._scheduled_work_column])
             row[self._scheduled_work_column] = swork
-            swork.name = row['task_name_mapped']
+            swork.model_name['granular_name'] = row['granular_name']
             swork.display_name = row['task_name']
             swork.volume = float(row['volume'])
             swork.start_end_time = Time(int(row['start'])), Time(int(row['finish']))
@@ -141,7 +147,7 @@ def from_scheduled_works(works: Iterable[ScheduledWork],
         """
         ordered_task_ids = order_nodes_by_start_time(works, wg) if wg else None
 
-        def sed(time1, time2, swork) -> tuple:
+        def sed(time1, time2) -> tuple:
             """
             Sorts times and calculates difference.
             :param time1: time 1.
@@ -151,19 +157,24 @@ def sed(time1, time2, swork) -> tuple:
             start, end = tuple(sorted((time1, time2)))
             return start, end, end - start
 
-        data_frame = [(i,                                                 # idx
-                       w.id,                                              # task_id
-                       w.display_name,                                    # task_name
-                       w.name,                                            # task_name_mapped
-                       w.contractor,                                      # contractor info
-                       w.cost,                                            # work cost
-                       w.volume,                                          # work volume
-                       w.volume_type,                                     # work volume type
-                       *sed(*(t.value for t in w.start_end_time), w),     # start, end, duration
-                       repr(dict((i.name, i.count) for i in w.workers)),  # workers
-                       w  # full ScheduledWork info
+        model_name_columns = _get_granular_name_columns(works)
+
+        def make_model_name_columns(swork: ScheduledWork) -> list[Any]:
+            return list(map(itemgetter(1), sorted(swork.model_name.items(), key=itemgetter(0))))
+
+        data_frame = [(i,                                                   # idx
+                       w.id,                                                # task_id
+                       w.display_name,                                      # task_name
+                       w.contractor,                                        # contractor info
+                       w.cost,                                              # work cost
+                       w.volume,                                            # work volume
+                       *sed(*(t.value for t in w.start_end_time)),          # start, end, duration
+                       repr(dict((i.name, i.count) for i in w.workers)),    # workers
+                       w,                                                   # full ScheduledWork info
+                       *make_model_name_columns(w),                         # model_name columns
                        ) for i, w in enumerate(works)]
-        data_frame = DataFrame.from_records(data_frame, columns=Schedule._columns)
+
+        data_frame = DataFrame.from_records(data_frame, columns=Schedule._columns + model_name_columns)
 
         data_frame = data_frame.set_index('idx', drop=False)
 
@@ -173,7 +184,7 @@ def sed(time1, time2, swork) -> tuple:
             data_frame = data_frame.sort_values(['task_id'])
             data_frame.task_id = data_frame.task_id.astype(str)
 
-        data_frame = data_frame.reindex(columns=Schedule._columns)
+        data_frame = data_frame.reindex(columns=Schedule._columns + model_name_columns)
         data_frame = data_frame.reset_index(drop=True)
 
         return Schedule(data_frame)

sampo/schemas/scheduled_work.py

@@ -38,11 +38,10 @@ def __init__(self,
                  materials: MaterialDelivery | None = None,
                  c_object: ConstructionObject | None = None):
         self.id = work_unit.id
-        self.name = work_unit.name
+        self.model_name = work_unit.model_name
         self.display_name = work_unit.display_name
         self.is_service_unit = work_unit.is_service_unit
         self.volume = work_unit.volume
-        self.volume_type = work_unit.volume_type
         self.priority = work_unit.priority
         self.start_end_time = start_end_time
         self.workers = workers if workers is not None else []
@@ -123,7 +122,7 @@ def duration(self) -> Time:
     def to_dict(self) -> dict[str, Any]:
         return {
             'task_id': self.id,
-            'task_name': self.name,
+            'task_name': self.model_name,
             'start': self.start_time.value,
             'finish': self.finish_time.value,
             'contractor_id': self.contractor,