Speed up owner match process

It's also completed in the notebook (at end).

eol_realign/notebooks/catalog-cargo-manifest_test.ipynb

@@ -354,7 +354,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 12,
+   "execution_count": 3,
    "metadata": {},
    "outputs": [
     {
@@ -368,7 +368,7 @@
        "      dtype='object')"
       ]
      },
-     "execution_count": 12,
+     "execution_count": 3,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -465,6 +465,360 @@
     "Only 6,219,674 of our 6,250,420 images were able to be pulled through the combined manifests."
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Align Owners to Manifest"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "owner_df = pd.read_csv(\"../data/media_manifest_missing_licenses_jul26_owners.csv\",\n",
+    "                       low_memory=False,\n",
+    "                       dtype={\"EOL content ID\": \"int64\", \"EOL page ID\": \"int64\"},)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "try merge then fill and concatenate."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "['filename_manifest', 'md5_combined_manifest', 'combined_id_manifest_redownload', 'eol_content_id', 'eol_page_id', 'medium_source_url', 'eol_full_size_copy_url', 'license_name', 'copyright_owner', 'expected_image_filename', 'source_0706', 'source_0726', 'source_1206', 'combined_id_full_manifest']\n",
+      "['EOL content ID', 'EOL page ID', 'Medium Source URL', 'EOL Full-Size Copy URL', 'License Name', 'Copyright Owner', 'license_link', 'title']\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(list(manifests.columns))\n",
+    "print(list(owner_df.columns))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "There are 12330062 entries in the combined manifest\n",
+      "There are 632043 entries in the owner match CSV\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(f\"There are {manifests.shape[0]} entries in the combined manifest\")\n",
+    "print(f\"There are {owner_df.shape[0]} entries in the owner match CSV\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "owner_df[\"combined_id_owner\"] = owner_df[\"EOL content ID\"].astype(str) + \"_\" + owner_df[\"EOL page ID\"].astype(str)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 630056 entries, 0 to 630055\n",
+      "Data columns (total 23 columns):\n",
+      " #   Column                           Non-Null Count   Dtype \n",
+      "---  ------                           --------------   ----- \n",
+      " 0   EOL content ID                   630056 non-null  int64 \n",
+      " 1   EOL page ID                      630056 non-null  int64 \n",
+      " 2   Medium Source URL                630056 non-null  object\n",
+      " 3   EOL Full-Size Copy URL           630056 non-null  object\n",
+      " 4   License Name                     630056 non-null  object\n",
+      " 5   Copyright Owner                  611829 non-null  object\n",
+      " 6   license_link                     619741 non-null  object\n",
+      " 7   title                            604447 non-null  object\n",
+      " 8   combined_id_owner                630056 non-null  object\n",
+      " 9   filename_manifest                630056 non-null  object\n",
+      " 10  md5_combined_manifest            630056 non-null  object\n",
+      " 11  combined_id_manifest_redownload  630056 non-null  object\n",
+      " 12  eol_content_id                   630056 non-null  int64 \n",
+      " 13  eol_page_id                      630056 non-null  int64 \n",
+      " 14  medium_source_url                630056 non-null  object\n",
+      " 15  eol_full_size_copy_url           630056 non-null  object\n",
+      " 16  license_name                     630056 non-null  object\n",
+      " 17  copyright_owner                  0 non-null       object\n",
+      " 18  expected_image_filename          630056 non-null  object\n",
+      " 19  source_0706                      630056 non-null  bool  \n",
+      " 20  source_0726                      630056 non-null  bool  \n",
+      " 21  source_1206                      630056 non-null  bool  \n",
+      " 22  combined_id_full_manifest        630056 non-null  object\n",
+      "dtypes: bool(3), int64(4), object(16)\n",
+      "memory usage: 97.9+ MB\n"
+     ]
+    }
+   ],
+   "source": [
+    "owners_manifest = pd.merge(owner_df,\n",
+    "                           manifests,\n",
+    "                           left_on=\"combined_id_owner\",\n",
+    "                           right_on = \"combined_id_full_manifest\",\n",
+    "                           how = \"inner\")\n",
+    "owners_manifest.info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 630056 entries, 0 to 630055\n",
+      "Data columns (total 23 columns):\n",
+      " #   Column                           Non-Null Count   Dtype \n",
+      "---  ------                           --------------   ----- \n",
+      " 0   EOL content ID                   630056 non-null  int64 \n",
+      " 1   EOL page ID                      630056 non-null  int64 \n",
+      " 2   Medium Source URL                630056 non-null  object\n",
+      " 3   EOL Full-Size Copy URL           630056 non-null  object\n",
+      " 4   License Name                     630056 non-null  object\n",
+      " 5   Copyright Owner                  611829 non-null  object\n",
+      " 6   license_link                     619741 non-null  object\n",
+      " 7   title                            604447 non-null  object\n",
+      " 8   combined_id_owner                630056 non-null  object\n",
+      " 9   filename_manifest                630056 non-null  object\n",
+      " 10  md5_combined_manifest            630056 non-null  object\n",
+      " 11  combined_id_manifest_redownload  630056 non-null  object\n",
+      " 12  eol_content_id                   630056 non-null  int64 \n",
+      " 13  eol_page_id                      630056 non-null  int64 \n",
+      " 14  medium_source_url                630056 non-null  object\n",
+      " 15  eol_full_size_copy_url           630056 non-null  object\n",
+      " 16  license_name                     630056 non-null  object\n",
+      " 17  copyright_owner                  611829 non-null  object\n",
+      " 18  expected_image_filename          630056 non-null  object\n",
+      " 19  source_0706                      630056 non-null  bool  \n",
+      " 20  source_0726                      630056 non-null  bool  \n",
+      " 21  source_1206                      630056 non-null  bool  \n",
+      " 22  combined_id_full_manifest        630056 non-null  object\n",
+      "dtypes: bool(3), int64(4), object(16)\n",
+      "memory usage: 97.9+ MB\n"
+     ]
+    }
+   ],
+   "source": [
+    "owners_manifest[\"copyright_owner\"] = owners_manifest[\"Copyright Owner\"]\n",
+    "owners_manifest.info(show_counts=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "Index(['filename_manifest', 'md5_combined_manifest',\n",
+       "       'combined_id_manifest_redownload', 'eol_content_id', 'eol_page_id',\n",
+       "       'medium_source_url', 'eol_full_size_copy_url', 'license_name',\n",
+       "       'copyright_owner', 'expected_image_filename', 'source_0706',\n",
+       "       'source_0726', 'source_1206', 'combined_id_full_manifest', 'title'],\n",
+       "      dtype='object')"
+      ]
+     },
+     "execution_count": 17,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "license_cols = list(manifests.columns)\n",
+    "license_cols.append(\"title\")\n",
+    "owners_manifest = owners_manifest[license_cols]\n",
+    "owners_manifest.columns"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "630056"
+      ]
+     },
+     "execution_count": 13,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "rematched_owner_ids = list(owners_manifest.combined_id_full_manifest.unique())\n",
+    "len(rematched_owner_ids)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(11700006, 14)"
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "un_matched_owner_manifest = manifests.loc[~manifests.combined_id_full_manifest.isin(rematched_owner_ids)]\n",
+    "un_matched_owner_manifest.shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/var/folders/nv/f0fq1p1n1_3b11x579py_0q80000gq/T/ipykernel_4207/1002247400.py:1: SettingWithCopyWarning: \n",
+      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
+      "Try using .loc[row_indexer,col_indexer] = value instead\n",
+      "\n",
+      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
+      "  un_matched_owner_manifest[\"title\"] = \"not provided\"\n"
+     ]
+    }
+   ],
+   "source": [
+    "un_matched_owner_manifest[\"title\"] = \"not provided\""
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 12330062 entries, 0 to 12330061\n",
+      "Data columns (total 15 columns):\n",
+      " #   Column                           Non-Null Count     Dtype \n",
+      "---  ------                           --------------     ----- \n",
+      " 0   filename_manifest                12330062 non-null  object\n",
+      " 1   md5_combined_manifest            12330062 non-null  object\n",
+      " 2   combined_id_manifest_redownload  12330062 non-null  object\n",
+      " 3   eol_content_id                   12330062 non-null  int64 \n",
+      " 4   eol_page_id                      12330062 non-null  int64 \n",
+      " 5   medium_source_url                12330062 non-null  object\n",
+      " 6   eol_full_size_copy_url           12330062 non-null  object\n",
+      " 7   license_name                     12330062 non-null  object\n",
+      " 8   copyright_owner                  11544885 non-null  object\n",
+      " 9   expected_image_filename          12330062 non-null  object\n",
+      " 10  source_0706                      12330062 non-null  bool  \n",
+      " 11  source_0726                      12330062 non-null  bool  \n",
+      " 12  source_1206                      12330062 non-null  bool  \n",
+      " 13  combined_id_full_manifest        12330062 non-null  object\n",
+      " 14  title                            12304453 non-null  object\n",
+      "dtypes: bool(3), int64(2), object(10)\n",
+      "memory usage: 1.1+ GB\n"
+     ]
+    }
+   ],
+   "source": [
+    "full_owner_manifest = pd.concat([un_matched_owner_manifest, owners_manifest], ignore_index=True)\n",
+    "full_owner_manifest.info(show_counts = True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Licenses still missing Copyright Owners: \n",
+      " license_name\n",
+      "cc-by-3.0                          262894\n",
+      "cc-publicdomain                    262803\n",
+      "cc-by-nc-sa-3.0                    122493\n",
+      "cc-by-sa-3.0                        93303\n",
+      "No known copyright restrictions     31025\n",
+      "cc-by-nc                            12619\n",
+      "cc-0-1.0                               32\n",
+      "cc-by-nc-2.0                            5\n",
+      "cc-by-nc-4.0                            2\n",
+      "cc-by-nc-sa                             1\n",
+      "Name: count, dtype: int64\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(\n",
+    "        \"Licenses still missing Copyright Owners: \\n\",\n",
+    "        full_owner_manifest.loc[\n",
+    "            full_owner_manifest[\"copyright_owner\"].isna(), \"license_name\"\n",
+    "        ].value_counts(),\n",
+    "    )"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "full_owner_manifest[\"copyright_owner\"].fillna(\"not provided\", inplace=True)\n",
+    "full_owner_manifest[\"title\"].fillna(\"not provided\", inplace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "full_owner_manifest.to_csv(\"../data/full_owner_manifest.csv\", index = False)"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,

eol_realign/scripts/make_licenses.py

@@ -57,7 +57,8 @@ def main(src_csv, dest_dir):
         if "license_name" not in df_cols:
             sys.exit("Source CSV does not have a column labeled License Name or license_name.")
         license_col = "license_name"
-    license_col = "License Name"
+    else:
+        license_col = "License Name"
 
     # Check filepath given by user
     dest_dir_path = Path(dest_dir)

eol_realign/scripts/match_owners.py

@@ -1,10 +1,9 @@
 import pandas as pd
-from tqdm import tqdm
 from pathlib import Path
 import argparse
 import sys
 
-# This file can be used on predicted-catalog or rarespecies-catalog by changing the CATALOG_PATH.
+# NOTE: column names are hard-coded to the files provided in MANIFEST_PATH & OWNER_PATH
 # It is intended to just add the extra licensing information to the combined manifest
 # Owner fix file: https://huggingface.co/datasets/imageomics/eol/blob/main/owners_info_fix/media_manifest_missing_licenses_jul26_owners.csv
 MANIFEST_PATH = "data/combined_manifest_with_checksums.csv"
@@ -24,35 +23,41 @@ def get_owners_titles(manifest, owners):
 
     Returns:
     --------
-    manifest - Media manifest DataFrame with missing owners resolved.
+    full_owner_manifest - Media manifest DataFrame with missing owners resolved.
+    num_matched - Number of records matched with the owner fix file.
     """
-    missing_owners = [
-        comb_id
-        for comb_id in list(
-            manifest.loc[manifest["copyright_owner"].isna(), "combined_id_full_manifest"]
-        )
-    ]
 
-    for comb_id in tqdm(missing_owners):
-        temp = owners.loc[owners.combined_id_owner == comb_id]
-        copyright_owner = str(temp["Copyright Owner"].values)
-        title = str(temp.title.values)
-        manifest.loc[manifest["combined_id_full_manifest"] == comb_id, "copyright_owner"] = copyright_owner
-        manifest.loc[manifest["combined_id_full_manifest"] == comb_id, "title"] = title
+    # First merge owners
+    owners_manifest = pd.merge(owners,
+                           manifest,
+                           left_on="combined_id_owner",
+                           right_on = "combined_id_full_manifest",
+                           how = "inner")
+    
+    # Set "copyright_owner" column to owners patched in owner CSV 
+    owners_manifest["copyright_owner"] = owners_manifest["Copyright Owner"]
 
-    # Print counts of licenses for which owner info was not resolved
-    print(
-        "Licenses still missing Copyright Owners: \n",
-        manifest.loc[
-            manifest["copyright_owner"].isna(), "License Name"
-        ].value_counts(),
-    )
+    # Preserve columns that match the combined manifest
+    license_cols = list(manifest.columns)
+    license_cols.append("title")
+    owners_manifest = owners_manifest[license_cols]
+
+    # Get matched owner IDs
+    rematched_owner_ids = list(owners_manifest.combined_id_full_manifest.unique())
+    num_matched = len(rematched_owner_ids)
+
+    # Reduce manifest to just records not matched in owner file
+    un_matched_owner_manifest = manifest.loc[~manifest.combined_id_full_manifest.isin(rematched_owner_ids)]
+    un_matched_owner_manifest["title"] = "not provided"
+
+    # Align the matched owners with the unmatched
+    full_owner_manifest = pd.concat([un_matched_owner_manifest, owners_manifest], ignore_index=True)
 
     # Fill null "copyright_owner" and "Title" values with "not provided"
-    manifest["copyright_owner"].fillna("not provided", inplace=True)
-    manifest["title"].fillna("not provided", inplace=True)
+    full_owner_manifest["copyright_owner"].fillna("not provided", inplace=True)
+    full_owner_manifest["title"].fillna("not provided", inplace=True)
 
-    return manifest
+    return full_owner_manifest, num_matched
 
 
 def update_owners(manifest, owners, filepath):
@@ -69,11 +74,20 @@ def update_owners(manifest, owners, filepath):
     # EOL content & page IDs read in as int64
     owners["combined_id_owner"] = owners["EOL content ID"].astype(str) + "_" + owners["EOL page ID"].astype(str)
 
-    updated_manifest = get_owners_titles(manifest, owners)
+    updated_manifest, num_matched = get_owners_titles(manifest, owners)
 
     # Save updated combined manifest file to chosen location
     updated_manifest.to_csv(filepath, index=False)
 
+    # Print counts of licenses for which owner info was not resolved
+    print(f"Of the {manifest.shape[0]} entries in manifest, {num_matched} were matched to the owner fix file.")
+    print(
+        "Licenses still missing Copyright Owners: \n",
+        manifest.loc[
+            manifest["copyright_owner"].isna(), "license_name"
+        ].value_counts(),
+    )
+
 
 def main(dest_dir):
     # Read in CSVs