From 96a2fbd2afc08a31b006b994d8252704438c8a07 Mon Sep 17 00:00:00 2001
From: Paurikova2 <107862249+Paurikova2@users.noreply.github.com>
Date: Mon, 24 Jun 2024 10:08:24 +0200
Subject: [PATCH] table status of tul before and after vanilla import

table status of tul before and after vanilla import - script
---
 src/dtb_difference.py | 42 ++++++++++++++++++++++++++++++++++++++++++
 src/pump/_db.py       |  9 +++++++--
 src/tul_settings.py   | 35 +++++++++++++++++++++++++++++++++++
 3 files changed, 84 insertions(+), 2 deletions(-)
 create mode 100644 src/dtb_difference.py
 create mode 100644 src/tul_settings.py

diff --git a/src/dtb_difference.py b/src/dtb_difference.py
new file mode 100644
index 0000000..ac2c58b
--- /dev/null
+++ b/src/dtb_difference.py
@@ -0,0 +1,42 @@
+import logging
+
+import settings
+import tul_settings
+from utils import init_logging, update_settings
+
+_logger = logging.getLogger()
+
+env = update_settings(settings.env, tul_settings.settings)
+init_logging(_logger, env["log_file"])
+
+import dspace  # noqa
+import pump
+
+
+def difference_dtb(old_dtb: dict, new_dtb: dict):
+    msg = ""
+    no_exist7 = ""
+    no_exist5 = ""
+    for name in sorted(old_dtb.keys()):
+        if name not in new_dtb:
+            no_exist7 += f"{name},"
+        else:
+            difference = int(new_dtb[name]) - int(old_dtb[name])
+            result = "surplus " if difference > 0 else (
+                "deficit " if difference < 0 else "")
+            msg += f"{name: >40}: {int(difference): >8d} {result}\n"
+        del new_dtb[name]
+    for name in sorted(new_dtb.keys()):
+        no_exist5 += f"{name},"
+    _logger.info(
+        f"\n{msg}Nonexistent tables in DSpace 7:\n\t{no_exist7}\nNonexistent tables in DSpace 5:\n\t{no_exist5}")
+    _logger.info(40 * "=")
+
+
+if __name__ == "__main__":
+    _logger.info("Loading repo objects")
+
+    _logger.info("Database difference:")
+    raw_db_7 = pump.db(env["db_dspace_7"])
+    raw_db_tul = pump.db(env["db_tul"])
+    difference_dtb(raw_db_tul.table_count(), raw_db_7.table_count())
diff --git a/src/pump/_db.py b/src/pump/_db.py
index 5372e0b..b3f65ed 100644
--- a/src/pump/_db.py
+++ b/src/pump/_db.py
@@ -127,13 +127,18 @@ def all_tables(self):
         return self.fetch_all(
             "SELECT table_name FROM information_schema.tables WHERE is_insertable_into = 'YES' AND table_schema = 'public'")
 
-    def status(self):
+    def table_count(self):
         d = {}
         tables = self.all_tables()
         for table in tables:
             name = table[0]
-            count = self.fetch_one(f"SELECT COUNT(*) FROM {name}")
+            # Use double quotes for table names because some of them are in uppercase.
+            count = self.fetch_one(f"SELECT COUNT(*) FROM \"{name}\"")
             d[name] = count
+        return d
+
+    def status(self):
+        d = self.table_count()
         zero = ""
         msg = ""
         for name in sorted(d.keys()):
diff --git a/src/tul_settings.py b/src/tul_settings.py
new file mode 100644
index 0000000..34b5e54
--- /dev/null
+++ b/src/tul_settings.py
@@ -0,0 +1,35 @@
+import os
+from datetime import datetime
+_this_dir = os.path.dirname(os.path.abspath(__file__))
+ts = datetime.now().strftime("%Y_%m_%d__%H.%M.%S")
+
+settings = {
+    "log_file": os.path.join(_this_dir, "../__logs", f"{ts}.txt"),
+
+    "resume_dir": "__temp/resume/",
+
+    "backend": {
+        "endpoint": "http://dev-5.pc:85/server/api/",
+        "user": "test@test.edu",
+        "password": "admin",
+        "authentication": True,
+    },
+
+    "db_dspace_7": {
+        # CLARIN-DSpace 7 database
+        "name": "dspace",
+        "host": "localhost",
+        # careful - NON standard port
+        "port": 5435,
+        "user": "dspace",
+        "password": "dspace",
+    },
+
+    "db_tul": {
+        "name": "tul",
+        "host": "localhost",
+        "user": "postgres",
+        "password": "dspace",
+        "port": 5432,
+    }
+}