From a6c2dc1b7d1634717c1ddff58b3bb70330d69db4 Mon Sep 17 00:00:00 2001
From: mdragaschnig <47592782+mdragaschnig@users.noreply.github.com>
Date: Wed, 22 Feb 2023 09:22:54 +0000
Subject: [PATCH] Revert "Add actor suggestion"

---
 README.md               | 15 ---------------
 src/tagpack/cli.py      | 31 -------------------------------
 src/tagpack/tagstore.py | 11 -----------
 3 files changed, 57 deletions(-)

diff --git a/README.md b/README.md
index 5bbb828..4a712e2 100644
--- a/README.md
+++ b/README.md
@@ -5,7 +5,6 @@
 This repository provides a command line tool for managing [GraphSense TagPacks](https://github.com/graphsense/graphsense-tagpacks/wiki/GraphSense-TagPacks). It can be used for
 
 1. [validating TagPacks against the TagPack schema](#validation)
-2. finding suitable actors for tags
 2. [validating ActorPacks against the ActorPack schema](#actorpack_validation)
 3. [handling taxonomies and concepts](#taxonomies)
 4. [ingesting TagPacks and related data into a TagStore](#tagstore)
@@ -31,20 +30,6 @@ Tagpacks are validated against the [tagpack schema](src/tagpack/conf/tagpack_sch
 
 Confidence settings are validated against a set of acceptable [confidence](src/tagpack/db/confidence.csv) values.
 
-## Actors for tags and TagPacks
-
-[Actors](https://github.com/graphsense/graphsense-tagpacks/wiki/Actors) are defined in a curated actor tagpack.
-
-It is encouraged to add actors to Tagpacks whenever possible.
-
-Finding suitable actors is supported by the tagpack-tool. For a specific lable, actor suggestions can be retrieved by calling
-
-    tagpack-tool tagpack suggest_actors <my_label>
-
-and if desired, the number of results can be restricted by adding the ``--max`` parameter
-
-    tagpack-tool tagpack suggest_actors --max 1 <my_label>
-
 ## Validate an ActorPack <a name="actorpack_validation"></a>
 
 Validate a single ActorPack file
diff --git a/src/tagpack/cli.py b/src/tagpack/cli.py
index cd8934e..84a9e2b 100644
--- a/src/tagpack/cli.py
+++ b/src/tagpack/cli.py
@@ -323,14 +323,6 @@ def validate_tagpack(args):
     )
 
 
-def suggest_actors(args):
-    print_line(f"Searching suitable actors for {args.label} in TagStore")
-
-    tagstore = TagStore(args.url, args.schema)
-    candidates = tagstore.find_actors_for(args.label, args.max)
-    print(f"Found {len(candidates)} matches: {candidates}")
-
-
 def insert_tagpack(args):
     t0 = time.time()
     print_line("TagPack insert starts")
@@ -865,29 +857,6 @@ def main():
     )
     ptp_i.set_defaults(func=insert_tagpack, url=def_url)
 
-    # parser for suggest_actor
-    ptp_actor = ptp.add_parser("suggest_actors", help="suggest actors ")
-    ptp_actor.add_argument(
-        "label",
-        nargs="?",
-        help="label string to find actor suggestions for",
-    )
-    ptp_actor.add_argument(
-        "--schema",
-        default=_DEFAULT_SCHEMA,
-        metavar="DB_SCHEMA",
-        help="PostgreSQL schema for tagpack tables",
-    )
-    ptp_actor.add_argument(
-        "-u", "--url", help="postgresql://user:password@db_host:port/database"
-    )
-    ptp_actor.add_argument(
-        "--max",
-        default=3,
-        help="Limits the number of results",
-    )
-    ptp_actor.set_defaults(func=suggest_actors, url=def_url)
-
     # parsers for actorpack command
     parser_ap = subparsers.add_parser("actorpack", help="actorpack commands")
 
diff --git a/src/tagpack/tagstore.py b/src/tagpack/tagstore.py
index 0f98929..eac2e84 100644
--- a/src/tagpack/tagstore.py
+++ b/src/tagpack/tagstore.py
@@ -179,17 +179,6 @@ def insert_actorpack(
 
         self.conn.commit()
 
-    def find_actors_for(self, label, max_results):
-        similarity = f"similarity(id, '{label}') > 0.2"
-
-        self.cursor.execute(
-            f"SELECT id FROM actor WHERE {similarity} "
-            f"ORDER BY {similarity} DESC "
-            f"LIMIT {max_results}"
-        )
-        matches = [x[0] for x in self.cursor.fetchall()]
-        return matches
-
     def low_quality_address_labels(self, th=0.25, currency="", category="") -> dict:
         """
         This function returns a list of addresses having a quality meassure