build_analytics/build_analytics/api_client.py

@@ -53,6 +53,7 @@ class APIclient():
         return self._parse_build(response.json())
 
     def __parse_build_node_stats(self, stats: Dict) -> BuildNodeStats:
+        logging.debug('raw json: %s', stats)
 
         keys = ['build_all', 'build_binaries', 'build_packages', 'build_srpm', 'build_node_task',
                 'cas_notarize_artifacts', 'cas_source_authenticate', 'git_checkout', 'upload']
@@ -66,11 +67,14 @@ class APIclient():
                         stats[k]['end_ts']+TZ_OFFSET) if stats[k]['end_ts'] else None)
             except KeyError:
                 params[k] = BuildStat()
-        return BuildNodeStats(**params)
+        build_node_stats = BuildNodeStats(**params)
+        logging.debug('BuildNodeStats: %s', build_node_stats)
+        return build_node_stats
 
     def __parse_web_node_stats(self, stats: Dict) -> WebNodeStats:
         keys = ['build_done', 'logs_processing', 'packages_processing']
         params = {}
+        logging.debug('raw json: %s', stats)
         for k in keys:
             try:
                 params[k] = BuildStat(
@@ -80,7 +84,9 @@ class APIclient():
                         stats[k]['end_ts']+TZ_OFFSET) if stats[k]['end_ts'] else None)
             except KeyError:
                 params[k] = BuildStat()
-        return WebNodeStats(**params)
+        web_node_stats = WebNodeStats(**params)
+        logging.debug('WebNodeStats %s', web_node_stats)
+        return web_node_stats
 
     def _parse_build_tasks(self, tasks_json: Dict, build_id: int) -> List[BuildTask]:
         result = []

build_analytics/build_analytics/const.py

@@ -2,7 +2,11 @@
 
 from enum import IntEnum
 
+# supported schema version
+DB_SCHEMA_VER = 1
 
+
+# ENUMS
 class ArchEnum(IntEnum):
     i686 = 0
     x86_64 = 1

build_analytics/build_analytics/db.py

@@ -1,5 +1,6 @@
 from datetime import datetime
-from typing import Union, Dict, List
+from typing import Union, Dict, List, Optional
+import logging
 
 import psycopg2
 
@@ -55,15 +56,18 @@ class DB():
                 '''
             cur.execute(sql, (stat.build_task_id, stat.stat_name_id,
                         stat.start_ts, stat.end_ts))
+            logging.debug('raw SQL query: %s', cur.query)
 
         # inserting build node stats
         for stat in build_node_stats:
+            logging.debug('BuildNodeStats: %s', stat)
             sql = '''
                     INSERT INTO build_node_stats(build_task_id, stat_name_id, start_ts, end_ts)
                     VALUES (%s, %s, %s, %s);
                 '''
             cur.execute(sql, (stat.build_task_id, stat.stat_name_id,
                         stat.start_ts, stat.end_ts))
+            logging.debug('raw SQL query: %s', cur.query)
 
         # commiting changes
         self.__conn.commit()
@@ -149,17 +153,31 @@ class DB():
                         end_ts = %s
                     WHERE build_task_id = %s;
                 '''
-            cur.execute(sql, (stat.start_ts, stat.end_ts))
+            cur.execute(sql, (stat.start_ts, stat.end_ts, build_task.id))
 
         # updating build_node_stats
         for stat in build_node_stats:
             sql = '''
                 UPDATE build_node_stats
                 SET start_ts = %s,
-                    end_ts = %s,
+                    end_ts = %s
                 WHERE build_task_id = %s;
                 '''
-            cur.execute(sql, (stat.start_ts, stat.end_ts))
+            cur.execute(sql, (stat.start_ts, stat.end_ts, build_task.id))
 
         # commiting changes
         self.__conn.commit()
+
+    def get_db_schema_version(self) -> Optional[int]:
+        sql = '''
+              SELECT *
+              FROM schema_version
+              LIMIT 1;
+              '''
+        cur = self.__conn.cursor()
+        cur.execute(sql)
+        val = cur.fetchone()
+        print(val)
+        if not val:
+            return None
+        return int(val[0])

build_analytics/build_analytics/extractor/extractor.py

@@ -4,7 +4,7 @@ import logging
 from typing import List, Dict
 
 from ..models.extractor_config import ExtractorConfig
-from ..models.enums import BuildTaskEnum
+from ..const import BuildTaskEnum
 from ..models.build import BuildTask
 from ..db import DB
 from ..api_client import APIclient
@@ -22,7 +22,7 @@ class Extractor:
         page_num = 1
         last_build_id = self.db.get_latest_build_id()
         if not last_build_id:
-            last_build_id = self.start_from
+            last_build_id = self.start_from - 1
         logging.info("last_build_id: %s", last_build_id)
         stop = False
 
@@ -49,7 +49,7 @@ class Extractor:
                         self.db.insert_buildtask(build_task.as_db_model(),
                                                  build_task.web_node_stats.as_db_model(
                                                      build_task.id),
-                                                 build_task.web_node_stats.as_db_model(
+                                                 build_task.build_node_stats.as_db_model(
                             build_task.id))
                     except Exception as error:  # pylint: disable=broad-except
                         logging.error('failed to insert build task %d: %s',

build_analytics/build_analytics/extractor/start.py

@@ -1,6 +1,7 @@
 from datetime import datetime, timedelta
 import logging
 from logging.handlers import RotatingFileHandler
+import sys
 import time
 
 import yaml
@@ -8,6 +9,7 @@ import yaml
 # pylint: disable=relative-beyond-top-level
 from ..api_client import APIclient
 from ..db import DB
+from ..const import DB_SCHEMA_VER
 from .extractor import Extractor
 from ..models.extractor_config import ExtractorConfig
 from ..models.db_config import DbConfig
@@ -46,16 +48,28 @@ def start(yml_path: str):
 
     # configuring logging
     logging.basicConfig(level=logging.INFO,
-                        format='%(asctime)s %(levelname)s <%(funcName)s> %(message)s',
+                        format='%(asctime)s %(levelname)s %(funcName)s() %(message)s',
                         handlers=[RotatingFileHandler(config.log_file,
                                                       maxBytes=10000000,
                                                       backupCount=3)])
 
+    # some pre-flight checks
+    db = DB(config.db_config)
+    cur_version = db.get_db_schema_version()
+    if not cur_version:
+        logging.error(
+            'Cant get db schema version. Make sure that schema_version exists')
+        sys.exit(1)
+    if cur_version != DB_SCHEMA_VER:
+        logging.error('unsupported DB schema: want %s, have %s',
+                      DB_SCHEMA_VER, cur_version)
+        sys.exit(1)
+
     while True:
         logging.info('Starting extraction proccess')
         api = APIclient(api_root=config.albs_url,
                         jwt=config.jwt, timeout=config.api_timeout)
-        db = DB(config.db_config)
+
         extractor = Extractor(config, api, db)
 
         logging.info('Starting builds insertion')

build_analytics/build_analytics/models/build_node_stats.py

@@ -5,7 +5,7 @@ from pydantic import BaseModel  # pylint: disable=no-name-in-module
 
 from .build_stat import BuildStat
 from .build_node_stat_db import BuildNodeStatDB
-from .enums import BuildNodeStatsEnum
+from ..const import BuildNodeStatsEnum
 
 
 class BuildNodeStats(BaseModel):

build_analytics/build_analytics/models/build_task.py

@@ -1,11 +1,11 @@
 from datetime import datetime
-from typing import Optional, Tuple
+from typing import Optional
 
 from pydantic import BaseModel  # pylint: disable=no-name-in-module
 
 from .build_task_db import BuildTaskDB
 from .build_node_stats import BuildNodeStats
-from .enums import ArchEnum
+from ..const import ArchEnum
 from .web_node_stats import WebNodeStats
 
 

build_analytics/build_analytics/models/web_node_stats.py

@@ -5,7 +5,7 @@ from pydantic import BaseModel  # pylint: disable=no-name-in-module
 
 from .build_stat import BuildStat
 from .web_node_stat_db import WebNodeStatDB
-from .enums import WebNodeStatsEnum
+from ..const import WebNodeStatsEnum
 
 
 class WebNodeStats(BaseModel):

build_analytics/config_default.yml

@@ -60,4 +60,4 @@ scrape_interval: 3600
 # build_id to start populating empty db with
 # required: false
 # default: 5808 (first build with correct metrics)
-start_from: 5808
+start_from:

build_analytics/migrations/1.sql

@@ -4,8 +4,8 @@ BEGIN;
 CREATE TABLE builds (
     id INTEGER PRIMARY KEY,
     url VARCHAR(50) NOT NULL,
-    created_at REAL NOT NULL,
+    created_at DOUBLE PRECISION NOT NULL,
-    finished_at REAL
+    finished_at DOUBLE PRECISION
 );
 
 
@@ -85,8 +85,8 @@ CREATE TABLE build_tasks (
     build_id INTEGER REFERENCES builds(id) ON DELETE CASCADE,
     arch_id INTEGER REFERENCES arch_enum(id) ON DELETE SET NULL,
     status_id INTEGER REFERENCES build_task_status_enum(id) ON DELETE SET NULL,
-    started_at REAL,
+    started_at DOUBLE PRECISION,
-    finished_at REAL
+    finished_at DOUBLE PRECISION
 );
 
 CREATE INDEX build_tasks_build_id
@@ -103,8 +103,8 @@ ON build_tasks(finished_at);
 CREATE TABLE web_node_stats (
     build_task_id INTEGER REFERENCES build_tasks(id) ON DELETE CASCADE,
     stat_name_id  INTEGER REFERENCES web_node_stats_enum(id) ON DELETE SET NULL,
-    start_ts REAL,
+    start_ts DOUBLE PRECISION,
-    end_ts REAL
+    end_ts DOUBLE PRECISION
 );
 
 CREATE INDEX web_node_stats_build_task_id
@@ -121,8 +121,8 @@ ON web_node_stats(end_ts);
 CREATE TABLE build_node_stats (
     build_task_id INTEGER REFERENCES build_tasks(id) ON DELETE CASCADE,
     stat_name_id  INTEGER REFERENCES build_node_stats_enum(id) ON DELETE SET NULL,
-    start_ts REAL,
+    start_ts DOUBLE PRECISION,
-    end_ts REAL
+    end_ts DOUBLE PRECISION
 );
 
 CREATE INDEX build_node_stats_build_task_id
@@ -140,8 +140,8 @@ CREATE TABLE sign_tasks (
     id INTEGER PRIMARY KEY,
     build_id INTEGER REFERENCES builds(id) ON DELETE CASCADE,
     buildtask_id INTEGER REFERENCES build_tasks(id) ON DELETE CASCADE,
-    started_at REAL,
+    started_at DOUBLE PRECISION,
-    finished_at REAL
+    finished_at DOUBLE PRECISION
 );
 
 CREATE INDEX sign_tasks_build_id
@@ -157,7 +157,6 @@ CREATE INDEX sign_tasks_finished_at
 ON sign_tasks(finished_at);
 
 
-
 -- schema_version
 CREATE TABLE schema_version (
     version INTEGER

build_analytics/releases.txt

@@ -2,4 +2,5 @@
 First version
 
 0.2.0
-New parameter start_from
+New parameter start_from
+moved to double persition for timestamps