gigafibre-fsm/scripts/migration/import_ticket_msgs.py

#!/usr/bin/env python3
"""
Import legacy ticket_msg → ERPNext Comment on Issue.

Maps: ticket_msg.ticket_id → Issue (via legacy_ticket_id)
      ticket_msg.staff_id → staff name for comment_by

Uses direct PostgreSQL INSERT for speed (784k+ messages).
Skips already-imported messages (checks by name pattern).

Run inside erpnext-backend-1:
  /home/frappe/frappe-bench/env/bin/python /home/frappe/frappe-bench/import_ticket_msgs.py
"""
import pymysql
import psycopg2
import uuid
from datetime import datetime, timezone

LEGACY = {"host": "10.100.80.100", "user": "facturation", "password": "VD67owoj",
          "database": "gestionclient", "connect_timeout": 30, "read_timeout": 600}
PG = {"host": "db", "port": 5432, "user": "postgres", "password": "123",
      "dbname": "_eb65bdc0c4b1b2d6"}

ADMIN = "Administrator"
BATCH_SIZE = 5000

def ts_to_dt(unix_ts):
    if not unix_ts or unix_ts <= 0:
        return None
    try:
        return datetime.fromtimestamp(int(unix_ts), tz=timezone.utc).strftime("%Y-%m-%d %H:%M:%S.%f")
    except (ValueError, OSError):
        return None

def log(msg):
    print(msg, flush=True)

def main():
    now = datetime.now(timezone.utc).strftime("%Y-%m-%d %H:%M:%S.%f")
    log("=== Import ticket_msg → Comment on Issue ===")

    # 1. Read legacy data
    log("Reading legacy staff...")
    mc = pymysql.connect(**LEGACY)
    cur = mc.cursor(pymysql.cursors.DictCursor)

    cur.execute("SELECT id, first_name, last_name, email FROM staff ORDER BY id")
    staff_list = cur.fetchall()
    staff_map = {}
    for s in staff_list:
        name = ((s.get("first_name") or "") + " " + (s.get("last_name") or "")).strip()
        staff_map[s["id"]] = {"name": name or "Staff #" + str(s["id"]), "email": s.get("email", "")}

    log("  {} staff loaded".format(len(staff_map)))

    # 2. Connect ERPNext PG
    log("Connecting to ERPNext PostgreSQL...")
    pg = psycopg2.connect(**PG)
    pgc = pg.cursor()

    # Build issue lookup: legacy_ticket_id → issue name
    pgc.execute('SELECT legacy_ticket_id, name FROM "tabIssue" WHERE legacy_ticket_id IS NOT NULL AND legacy_ticket_id > 0')
    issue_map = {r[0]: r[1] for r in pgc.fetchall()}
    log("  {} issues mapped".format(len(issue_map)))

    # Check existing imported comments (by name pattern TMSG-)
    pgc.execute("""SELECT name FROM "tabComment" WHERE name LIKE 'TMSG-%'""")
    existing = set(r[0] for r in pgc.fetchall())
    log("  {} existing TMSG comments (will skip)".format(len(existing)))

    # 3. Read and import messages in batches
    log("Reading ticket_msg from legacy (streaming)...")
    cur_stream = mc.cursor(pymysql.cursors.SSDictCursor)
    cur_stream.execute("""SELECT id, ticket_id, staff_id, msg, date_orig, public, important
                          FROM ticket_msg ORDER BY ticket_id, id""")

    ok = skip_no_issue = skip_existing = skip_empty = err = 0
    batch = []
    total_read = 0

    for row in cur_stream:
        total_read += 1
        tid = row["ticket_id"]
        mid = row["id"]
        msg_name = "TMSG-{}".format(mid)

        # Skip if already imported
        if msg_name in existing:
            skip_existing += 1
            continue

        # Skip if no matching issue
        issue_name = issue_map.get(tid)
        if not issue_name:
            skip_no_issue += 1
            continue

        # Skip empty messages
        msg_text = row.get("msg") or ""
        if not msg_text.strip():
            skip_empty += 1
            continue

        staff = staff_map.get(row.get("staff_id"), {"name": "Système", "email": ""})
        msg_date = ts_to_dt(row.get("date_orig")) or now

        batch.append((
            msg_name,           # name
            msg_date,           # creation
            msg_date,           # modified
            ADMIN,              # modified_by
            staff["email"] or ADMIN,  # owner
            0,                  # docstatus
            0,                  # idx
            "Comment",          # comment_type
            staff["email"],     # comment_email
            "",                 # subject
            staff["name"],      # comment_by
            0,                  # published
            1,                  # seen
            "Issue",            # reference_doctype
            issue_name,         # reference_name
            ADMIN,              # reference_owner
            msg_text,           # content
        ))

        if len(batch) >= BATCH_SIZE:
            try:
                _insert_batch(pgc, batch)
                pg.commit()
                ok += len(batch)
            except Exception as e:
                pg.rollback()
                # Fallback: row by row
                for b in batch:
                    try:
                        _insert_batch(pgc, [b])
                        pg.commit()
                        ok += 1
                    except Exception:
                        pg.rollback()
                        err += 1
            batch = []

            if ok % 50000 == 0:
                log("  read={} ok={} skip_issue={} skip_dup={} skip_empty={} err={}".format(
                    total_read, ok, skip_no_issue, skip_existing, skip_empty, err))

    # Final batch
    if batch:
        try:
            _insert_batch(pgc, batch)
            pg.commit()
            ok += len(batch)
        except Exception:
            pg.rollback()
            for b in batch:
                try:
                    _insert_batch(pgc, [b])
                    pg.commit()
                    ok += 1
                except Exception:
                    pg.rollback()
                    err += 1

    cur_stream.close()
    mc.close()
    pg.close()

    log("")
    log("=" * 60)
    log("Total read:      {}".format(total_read))
    log("Imported:        {}".format(ok))
    log("Skip (no issue): {}".format(skip_no_issue))
    log("Skip (existing): {}".format(skip_existing))
    log("Skip (empty):    {}".format(skip_empty))
    log("Errors:          {}".format(err))
    log("=" * 60)


def _insert_batch(pgc, rows):
    """Insert batch of Comment rows."""
    args = ",".join(
        pgc.mogrify("(%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s)", r).decode()
        for r in rows
    )
    pgc.execute("""
        INSERT INTO "tabComment" (
            name, creation, modified, modified_by, owner, docstatus, idx,
            comment_type, comment_email, subject, comment_by, published, seen,
            reference_doctype, reference_name, reference_owner, content
        ) VALUES """ + args + """ ON CONFLICT (name) DO NOTHING""")


if __name__ == "__main__":
    main()