history: use tables for history and totals storage

Signed-off-by:  Eric Callahan <arksine.code@gmail.com>
This commit is contained in:
Eric Callahan 2024-05-08 13:58:08 -04:00
parent 80c762074f
commit 3f0d20ed8c
1 changed files with 352 additions and 191 deletions

View File

@ -1,5 +1,7 @@
# History cache for printer jobs
#
# Copyright (C) 2024 Eric Callahan <arksine.code@gmail.com>
#
# This file may be distributed under the terms of the GNU GPLv3 license.
from __future__ import annotations
@ -10,7 +12,8 @@ from ..common import (
JobEvent,
RequestType,
HistoryFieldData,
FieldTracker
FieldTracker,
SqlTableDefinition
)
# Annotation imports
@ -20,7 +23,8 @@ from typing import (
Union,
Optional,
Dict,
List
List,
Tuple
)
if TYPE_CHECKING:
@ -29,13 +33,10 @@ if TYPE_CHECKING:
from .database import MoonrakerDatabase as DBComp
from .job_state import JobState
from .file_manager.file_manager import FileManager
from .database import DBProviderWrapper
Totals = Dict[str, Union[float, int]]
AuxTotals = List[Dict[str, Any]]
HIST_NAMESPACE = "history"
HIST_VERSION = 1
MAX_JOBS = 10000
BASE_TOTALS = {
"total_jobs": 0,
"total_time": 0.,
@ -44,20 +45,142 @@ BASE_TOTALS = {
"longest_job": 0.,
"longest_print": 0.
}
HIST_TABLE = "job_history"
TOTALS_TABLE = "job_totals"
def _create_totals_list(
job_totals: Dict[str, Any],
aux_totals: List[Dict[str, Any]],
instance: str = "default"
) -> List[Tuple[str, str, Any, Any, str]]:
"""
Returns a list of Tuples formatted for SQL Database insertion.
Fields of each tuple are in the following order:
provider, field, maximum, total, instance_id
"""
totals_list: List[Tuple[str, str, Any, Any, str]] = []
for key, value in job_totals.items():
total = value if key.startswith("total_") else None
maximum = value if total is None else None
totals_list.append(("history", key, maximum, total, instance))
for item in aux_totals:
totals_list.append(
(
item["provider"],
item["field"],
item["maximum"],
item["total"],
instance
)
)
return totals_list
class TotalsSqlDefinition(SqlTableDefinition):
name = TOTALS_TABLE
prototype = (
f"""
{TOTALS_TABLE} (
provider TEXT NOT NULL,
field TEXT NOT NULL,
maximum REAL,
total REAL,
instance_id TEXT NOT NULL,
PRIMARY KEY (provider, field, instance_id)
)
"""
)
version = 1
def migrate(self, last_version: int, db_provider: DBProviderWrapper) -> None:
if last_version == 0:
# Migrate from "moonraker" namespace to a table
logging.info("Migrating history totals from moonraker namespace...")
hist_ns: Dict[str, Any] = db_provider.get_item("moonraker", "history", {})
job_totals: Dict[str, Any] = hist_ns.get("job_totals", BASE_TOTALS)
aux_totals: List[Dict[str, Any]] = hist_ns.get("aux_totals", [])
totals_list = _create_totals_list(job_totals, aux_totals)
sql_conn = db_provider.connection
with sql_conn:
sql_conn.executemany(
f"INSERT OR IGNORE INTO {TOTALS_TABLE} VALUES(?, ?, ?, ?, ?)",
totals_list
)
try:
db_provider.delete_item("moonraker", "history")
except Exception:
pass
class HistorySqlDefinition(SqlTableDefinition):
name = HIST_TABLE
prototype = (
f"""
{HIST_TABLE} (
job_id INTEGER PRIMARY KEY ASC,
user TEXT NOT NULL,
filename TEXT,
status TEXT NOT NULL,
start_time REAL NOT NULL,
end_time REAL,
print_duration REAL NOT NULL,
total_duration REAL NOT NULL,
filament_used REAL NOT NULL,
metadata pyjson,
auxiliary_data pyjson NOT NULL,
instance_id TEXT NOT NULL
)
"""
)
version = 1
def migrate(self, last_version: int, db_provider: DBProviderWrapper) -> None:
if last_version == 0:
conn = db_provider.connection
for batch in db_provider.iter_namespace("history", 1000):
conv_vals: List[Tuple[Any, ...]] = []
entry: Dict[str, Any]
for entry in batch.values():
try:
conv_vals.append(
(
None,
entry.get("user", "No User"),
entry["filename"],
entry["status"],
entry["start_time"],
entry["end_time"],
entry["print_duration"],
entry["total_duration"],
entry["filament_used"],
entry["metadata"],
entry.get("auxiliary_data", []),
"default"
)
)
except KeyError:
continue
if not conv_vals:
continue
placeholders = ",".join("?" * len(conv_vals[0]))
with conn:
conn.executemany(
f"INSERT INTO {HIST_TABLE} VALUES({placeholders})",
conv_vals
)
db_provider.wipe_local_namespace("history")
class History:
def __init__(self, config: ConfigHelper) -> None:
self.server = config.get_server()
self.file_manager: FileManager = self.server.lookup_component(
'file_manager')
self.file_manager: FileManager = self.server.lookup_component('file_manager')
self.request_lock = Lock()
FieldTracker.class_init(self)
self.auxiliary_fields: List[HistoryFieldData] = []
database: DBComp = self.server.lookup_component("database")
hist_info: Dict[str, Any]
hist_info = database.get_item("moonraker", "history", {}).result()
self.job_totals: Totals = hist_info.get("job_totals", dict(BASE_TOTALS))
self.aux_totals: AuxTotals = hist_info.get("aux_totals", [])
self.history_table = database.register_table(HistorySqlDefinition())
self.totals_table = database.register_table(TotalsSqlDefinition())
self.job_totals: Totals = dict(BASE_TOTALS)
self.aux_totals: AuxTotals = []
self.server.register_event_handler(
"server:klippy_disconnect", self._handle_disconnect)
@ -84,41 +207,43 @@ class History:
self._handle_job_total_reset
)
database.register_local_namespace(HIST_NAMESPACE)
self.history_ns = database.wrap_namespace(HIST_NAMESPACE,
parse_keys=False)
self.current_job: Optional[PrinterJob] = None
self.current_job_id: Optional[str] = None
self.current_job_id: Optional[int] = None
self.job_user: str = "No User"
self.job_paused: bool = False
self.next_job_id: int = 0
self.cached_job_ids = self.history_ns.keys().result()
if self.cached_job_ids:
self.next_job_id = int(self.cached_job_ids[-1], 16) + 1
async def component_init(self) -> None:
# Check for interupted jobs. If this is the first time, check
# the entire database. Otherwise only check the last 20 jobs.
interrupted_jobs: Dict[str, Any] = {}
database: DBComp = self.server.lookup_component("database")
version: int = await database.get_item("moonraker", "history.version", 0)
if version != HIST_VERSION:
await database.insert_item("moonraker", "history.version", HIST_VERSION)
job_ids = self.cached_job_ids if version < 1 else self.cached_job_ids[-20:]
jobs: Dict[str, Dict[str, Any]]
jobs = await self.history_ns.get_batch(job_ids)
for jid, job_data in jobs.items():
if job_data.get("status", "") == "in_progress":
job_data["status"] = "interrupted"
interrupted_jobs[jid] = job_data
# Popluate totals
valid_aux_totals = [
(item.provider, item.name) for item in self.auxiliary_fields
if item.has_totals()
]
cursor = await self.totals_table.execute(f"SELECT * from {TOTALS_TABLE}")
await cursor.set_arraysize(200)
for row in await cursor.fetchall():
provider, field, maximum, total, _ = tuple(row)
if provider == "history":
self.job_totals[field] = total if maximum is None else maximum
elif (provider, field) in valid_aux_totals:
item = dict(row)
item.pop("instance_id", None)
self.aux_totals.append(item)
# Check for interupted jobs
cursor = await self.history_table.execute(
f"SELECT job_id FROM {HIST_TABLE} WHERE status = 'in_progress'"
)
interrupted_jobs: List[int] = [row[0] for row in await cursor.fetchall()]
if interrupted_jobs:
async with self.history_table as tx:
await tx.execute(
f"UPDATE {HIST_TABLE} SET status = 'interrupted' "
"WHERE status = 'in_progress'"
)
self.server.add_log_rollover_item(
"interrupted_history",
"The following jobs were detected as interrupted: "
f"{list(interrupted_jobs.keys())}"
f"{interrupted_jobs}"
)
await self.history_ns.insert_batch(interrupted_jobs)
async def _handle_job_request(self,
web_request: WebRequest
@ -127,24 +252,37 @@ class History:
req_type = web_request.get_request_type()
if req_type == RequestType.GET:
job_id = web_request.get_str("uid")
if job_id not in self.cached_job_ids:
cursor = await self.history_table.execute(
f"SELECT * FROM {HIST_TABLE} WHERE job_id = ?", (int(job_id, 16),)
)
result = await cursor.fetchone()
if result is None:
raise self.server.error(f"Invalid job uid: {job_id}", 404)
job = await self.history_ns[job_id]
job = dict(result)
return {"job": self._prep_requested_job(job, job_id)}
if req_type == RequestType.DELETE:
all = web_request.get_boolean("all", False)
if all:
deljobs = self.cached_job_ids
self.history_ns.clear()
self.cached_job_ids = []
self.next_job_id = 0
cursor = await self.history_table.execute(
f"SELECT job_id FROM {HIST_TABLE} WHERE instance_id = ?",
("default",)
)
await cursor.set_arraysize(1000)
deljobs = [f"{row[0]:06X}" for row in await cursor.fetchall()]
async with self.history_table as tx:
await tx.execute(
f"DELETE FROM {HIST_TABLE} WHERE instance_id = ?",
("default",)
)
return {'deleted_jobs': deljobs}
job_id = web_request.get_str("uid")
if job_id not in self.cached_job_ids:
async with self.history_table as tx:
cursor = await tx.execute(
f"DELETE FROM {HIST_TABLE} WHERE job_id = ?", (int(job_id, 16),)
)
if cursor.rowcount < 1:
raise self.server.error(f"Invalid job uid: {job_id}", 404)
self.delete_job(job_id)
return {'deleted_jobs': [job_id]}
raise self.server.error("Invalid Request Method")
@ -152,59 +290,36 @@ class History:
web_request: WebRequest
) -> Dict[str, Any]:
async with self.request_lock:
i = 0
count = 0
end_num = len(self.cached_job_ids)
jobs: List[Dict[str, Any]] = []
start_num = 0
before = web_request.get_float("before", -1)
since = web_request.get_float("since", -1)
limit = web_request.get_int("limit", 50)
start = web_request.get_int("start", 0)
order = web_request.get_str("order", "desc")
order = web_request.get_str("order", "desc").upper()
if order not in ["asc", "desc"]:
if order not in ["ASC", "DESC"]:
raise self.server.error(f"Invalid `order` value: {order}", 400)
reverse_order = (order == "desc")
# cached jobs is asc order, find lower and upper boundary
if since != -1:
while start_num < end_num:
job_id = self.cached_job_ids[start_num]
job: Dict[str, Any] = await self.history_ns[job_id]
if job['start_time'] > since:
break
start_num += 1
# Build SQL Select Statement
values: List[Any] = ["default"]
sql_statement = f"SELECT * FROM {HIST_TABLE} WHERE instance_id = ?"
if before != -1:
while end_num > 0:
job_id = self.cached_job_ids[end_num-1]
job = await self.history_ns[job_id]
if job['end_time'] < before:
break
end_num -= 1
if start_num >= end_num or end_num == 0:
return {"count": 0, "jobs": []}
i = start
count = end_num - start_num
if limit == 0:
limit = MAX_JOBS
while i < count and len(jobs) < limit:
if reverse_order:
job_id = self.cached_job_ids[end_num - i - 1]
else:
job_id = self.cached_job_ids[start_num + i]
job = await self.history_ns[job_id]
sql_statement += " and end_time < ?"
values.append(before)
if since != -1:
sql_statement += " and start_time > ?"
values.append(since)
sql_statement += f" ORDER BY job_id {order}"
if limit > 0:
sql_statement += " LIMIT ? OFFSET ?"
values.append(limit)
values.append(start)
cursor = await self.history_table.execute(sql_statement, values)
await cursor.set_arraysize(1000)
jobs: List[Dict[str, Any]] = []
for row in await cursor.fetchall():
job = dict(row)
job_id = f"{row['job_id']:06X}"
jobs.append(self._prep_requested_job(job, job_id))
i += 1
return {"count": count, "jobs": jobs}
return {"count": len(jobs), "jobs": jobs}
async def _handle_job_totals(
self, web_request: WebRequest
@ -223,15 +338,20 @@ class History:
self.job_totals = dict(BASE_TOTALS)
last_aux_totals = self.aux_totals
self._update_aux_totals(reset=True)
database: DBComp = self.server.lookup_component("database")
await database.insert_item("moonraker", "history.job_totals", self.job_totals)
await database.insert_item("moonraker", "history.aux_totals", self.aux_totals)
totals_list = _create_totals_list(self.job_totals, self.aux_totals)
async with self.totals_table as tx:
await tx.execute(
f"DELETE FROM {TOTALS_TABLE} WHERE instance_id = ?", ("default",)
)
await tx.executemany(
f"INSERT INTO {TOTALS_TABLE} VALUES(?, ?, ?, ?, ?)", totals_list
)
return {
"last_totals": last_totals,
"last_auxiliary_totals": last_aux_totals
}
def _on_job_state_changed(
async def _on_job_state_changed(
self,
job_event: JobEvent,
prev_stats: Dict[str, Any],
@ -241,16 +361,16 @@ class History:
if job_event == JobEvent.STARTED:
if self.current_job is not None:
# Finish with the previous state
self.finish_job("cancelled", prev_stats)
self.add_job(PrinterJob(new_stats))
await self.finish_job("cancelled", prev_stats)
await self.add_job(PrinterJob(new_stats))
elif job_event == JobEvent.COMPLETE:
self.finish_job("completed", new_stats)
await self.finish_job("completed", new_stats)
elif job_event == JobEvent.ERROR:
self.finish_job("error", new_stats)
await self.finish_job("error", new_stats)
elif job_event in (JobEvent.CANCELLED, JobEvent.STANDBY):
# Cancel on "standby" for backward compatibility with
# `CLEAR_PAUSE/SDCARD_RESET_FILE` workflow
self.finish_job("cancelled", prev_stats)
await self.finish_job("cancelled", prev_stats)
def _on_job_requested(self, user: Optional[Dict[str, Any]]) -> None:
username = (user or {}).get("username", "No User")
@ -258,92 +378,117 @@ class History:
if self.current_job is not None:
self.current_job.user = username
def _handle_shutdown(self) -> None:
async def _handle_shutdown(self) -> None:
jstate: JobState = self.server.lookup_component("job_state")
last_ps = jstate.get_last_stats()
self.finish_job("klippy_shutdown", last_ps)
await self.finish_job("klippy_shutdown", last_ps)
def _handle_disconnect(self) -> None:
async def _handle_disconnect(self) -> None:
jstate: JobState = self.server.lookup_component("job_state")
last_ps = jstate.get_last_stats()
self.finish_job("klippy_disconnect", last_ps)
await self.finish_job("klippy_disconnect", last_ps)
def add_job(self, job: PrinterJob) -> None:
if len(self.cached_job_ids) >= MAX_JOBS:
self.delete_job(self.cached_job_ids[0])
job_id = f"{self.next_job_id:06X}"
self.current_job = job
self.current_job_id = job_id
self.current_job.user = self.job_user
self.grab_job_metadata()
for field in self.auxiliary_fields:
field.tracker.reset()
self.current_job.set_aux_data(self.auxiliary_fields)
self.history_ns[job_id] = job.get_stats()
self.cached_job_ids.append(job_id)
self.next_job_id += 1
logging.debug(
f"History Job Added - Id: {job_id}, File: {job.filename}"
)
self.send_history_event("added")
async def add_job(self, job: PrinterJob) -> None:
async with self.request_lock:
self.current_job = job
self.current_job_id = None
self.current_job.user = self.job_user
self.grab_job_metadata()
for field in self.auxiliary_fields:
field.tracker.reset()
self.current_job.set_aux_data(self.auxiliary_fields)
new_id = await self.save_job(job, None)
if new_id is None:
logging.info(f"Error saving job, filename '{job.filename}'")
return
self.current_job_id = new_id
job_id = f"{new_id:06X}"
self.update_metadata(job_id)
logging.debug(
f"History Job Added - Id: {job_id}, File: {job.filename}"
)
self.send_history_event("added")
def delete_job(self, job_id: Union[int, str]) -> None:
if isinstance(job_id, int):
job_id = f"{job_id:06X}"
async def save_job(self, job: PrinterJob, job_id: Optional[int]) -> Optional[int]:
values: List[Any] = [
job_id,
job.user,
job.filename,
job.status,
job.start_time,
job.end_time,
job.print_duration,
job.total_duration,
job.filament_used,
job.metadata,
job.auxiliary_data,
"default"
]
placeholders = ",".join("?" * len(values))
async with self.history_table as tx:
cursor = await tx.execute(
f"REPLACE INTO {HIST_TABLE} VALUES({placeholders})", values
)
return cursor.lastrowid
if job_id in self.cached_job_ids:
del self.history_ns[job_id]
self.cached_job_ids.remove(job_id)
async def delete_job(self, job_id: Union[int, str]) -> None:
if isinstance(job_id, str):
job_id = int(job_id, 16)
async with self.history_table as tx:
tx.execute(
f"DELETE FROM {HIST_TABLE} WHERE job_id = ?", (job_id,)
)
def finish_job(self, status: str, pstats: Dict[str, Any]) -> None:
if self.current_job is None:
return
cj = self.current_job
if (
pstats.get('filename') != cj.get('filename') or
pstats.get('total_duration', 0.) < cj.get('total_duration')
):
# Print stats have been reset, do not update this job with them
pstats = {}
async def finish_job(self, status: str, pstats: Dict[str, Any]) -> None:
async with self.request_lock:
if self.current_job is None or self.current_job_id is None:
self._reset_current_job()
return
if (
pstats.get('filename') != self.current_job.filename or
pstats.get('total_duration', 0.) < self.current_job.total_duration
):
# Print stats have been reset, do not update this job with them
pstats = {}
self.current_job.user = self.job_user
self.current_job.finish(status, pstats)
# Regrab metadata incase metadata wasn't parsed yet due to file upload
self.grab_job_metadata()
self.current_job.set_aux_data(self.auxiliary_fields)
job_id = f"{self.current_job_id:06X}"
await self.save_job(self.current_job, self.current_job_id)
self.update_metadata(job_id)
await self._update_job_totals()
logging.debug(
f"History Job Finished - Id: {job_id}, "
f"File: {self.current_job.filename}, "
f"Status: {status}"
)
self.send_history_event("finished")
self._reset_current_job()
self.current_job.user = self.job_user
self.current_job.finish(status, pstats)
# Regrab metadata incase metadata wasn't parsed yet due to file upload
self.grab_job_metadata()
self.current_job.set_aux_data(self.auxiliary_fields)
self.save_current_job()
self._update_job_totals()
logging.debug(
f"History Job Finished - Id: {self.current_job_id}, "
f"File: {self.current_job.filename}, "
f"Status: {status}"
)
self.send_history_event("finished")
def _reset_current_job(self) -> None:
self.current_job = None
self.current_job_id = None
self.job_user = "No User"
async def get_job(self,
job_id: Union[int, str]
) -> Optional[Dict[str, Any]]:
if isinstance(job_id, int):
job_id = f"{job_id:06X}"
return await self.history_ns.get(job_id, None)
async def get_job(
self, job_id: Union[int, str]
) -> Optional[Dict[str, Any]]:
if isinstance(job_id, str):
job_id = int(job_id, 16)
cursor = await self.history_table.execute(
f"SELECT * FROM {HIST_TABLE} WHERE job_id = ?", (job_id,)
)
result = await cursor.fetchone()
return dict(result) if result is not None else result
def grab_job_metadata(self) -> None:
if self.current_job is None:
return
filename: str = self.current_job.get("filename")
filename: str = self.current_job.filename
mdst = self.file_manager.get_metadata_storage()
metadata: Dict[str, Any] = mdst.get(filename, {})
if metadata:
# Add the start time and job id to the
# persistent metadata storage
metadata.update({
'print_start_time': self.current_job.get('start_time'),
'job_id': self.current_job_id
})
mdst.insert(filename, metadata.copy())
# We don't need to store these fields in the
# job metadata, as they are redundant
metadata.pop('print_start_time', None)
@ -352,14 +497,24 @@ class History:
thumb: Dict[str, Any]
for thumb in metadata['thumbnails']:
thumb.pop('data', None)
self.current_job.set("metadata", metadata)
self.current_job.metadata = metadata
def save_current_job(self) -> None:
if self.current_job is None or self.current_job_id is None:
def update_metadata(self, job_id: str) -> None:
if self.current_job is None:
return
self.history_ns[self.current_job_id] = self.current_job.get_stats()
mdst = self.file_manager.get_metadata_storage()
filename: str = self.current_job.filename
metadata: Dict[str, Any] = mdst.get(filename, {})
if metadata:
# Add the start time and job id to the
# persistent metadata storage
metadata.update({
'print_start_time': self.current_job.get('start_time'),
'job_id': job_id
})
mdst.insert(filename, metadata)
def _update_job_totals(self) -> None:
async def _update_job_totals(self) -> None:
if self.current_job is None:
return
job = self.current_job
@ -370,9 +525,11 @@ class History:
self._maximize_total("longest_job", job.total_duration)
self._maximize_total("longest_print", job.print_duration)
self._update_aux_totals()
database: DBComp = self.server.lookup_component("database")
database.insert_item("moonraker", "history.job_totals", self.job_totals)
database.insert_item("moonraker", "history.aux_totals", self.aux_totals)
totals_list = _create_totals_list(self.job_totals, self.aux_totals)
async with self.totals_table as tx:
await tx.executemany(
f"REPLACE INTO {TOTALS_TABLE} VALUES(?, ?, ?, ?, ?)", totals_list
)
def _accumulate_total(self, field: str, val: Union[int, float]) -> None:
self.job_totals[field] += val
@ -391,23 +548,25 @@ class History:
def send_history_event(self, evt_action: str) -> None:
if self.current_job is None or self.current_job_id is None:
return
job = self._prep_requested_job(
self.current_job.get_stats(), self.current_job_id)
self.server.send_event("history:history_changed",
{'action': evt_action, 'job': job})
def _prep_requested_job(self,
job: Dict[str, Any],
job_id: str
) -> Dict[str, Any]:
mtime = job.get("metadata", {}).get("modified", None)
job['job_id'] = job_id
job['exists'] = self.file_manager.check_file_exists(
"gcodes", job['filename'], mtime
job_id = f"{self.current_job_id:06X}"
job = self._prep_requested_job(self.current_job.get_stats(), job_id)
self.server.send_event(
"history:history_changed", {'action': evt_action, 'job': job}
)
def _prep_requested_job(
self, job: Dict[str, Any], job_id: str
) -> Dict[str, Any]:
fm = self.file_manager
mtime = job.get("metadata", {}).get("modified", None)
job["exists"] = fm.check_file_exists("gcodes", job['filename'], mtime)
job["job_id"] = job_id
job.pop("instance_id", None)
return job
def register_auxiliary_field(self, new_field: HistoryFieldData) -> None:
if new_field.provider == "history":
raise self.server.error("Provider name 'history' is reserved")
for field in self.auxiliary_fields:
if field == new_field:
raise self.server.error(
@ -421,17 +580,19 @@ class History:
return False
return not self.job_paused if check_paused else True
def on_exit(self) -> None:
async def on_exit(self) -> None:
if self.current_job is None:
return
jstate: JobState = self.server.lookup_component("job_state")
last_ps = jstate.get_last_stats()
self.finish_job("server_exit", last_ps)
await self.finish_job("server_exit", last_ps)
class PrinterJob:
def __init__(self, data: Dict[str, Any] = {}) -> None:
self.end_time: Optional[float] = None
self.filament_used: float = 0
self.filename: Optional[str] = None
self.metadata: Optional[Dict[str, Any]] = None
self.filename: str = ""
self.metadata: Dict[str, Any] = {}
self.print_duration: float = 0.
self.status: str = "in_progress"
self.start_time = time.time()