Remove "-" causing problems with dataset viewer (#8)

- update lfs tracking in gitattributes (929680d0ec74222a0d89f842a6cdcad422434b69)
- Replace '-' in 'Collected_by' column to fix dataset viewer rendering (03f93bf7afdab95e7c67e49f13d4a8a6f6f205ee)
- Notebook finding and fixing the '-' in 'Collected_by' column of full and heliconius CSVs (d65770f7a85dae010c6ebfc9592567f472c25ef6)

.gitattributes

@@ -55,8 +55,11 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.webp filter=lfs diff=lfs merge=lfs -text
 
 # Large CSVs
-Jiggins_Zenodo_Img_Master.csv filter=lfs diff=lfs merge=lfs -text
-Jiggins_Heliconius_Master.csv filter=lfs diff=lfs merge=lfs -text
+img_master.csv filter=lfs diff=lfs merge=lfs -text
+Heliconius_img_master.csv filter=lfs diff=lfs merge=lfs -text
 metadata/Jiggins_Zenodo_Master.csv filter=lfs diff=lfs merge=lfs -text
 deduplication_process/Jiggins_Zenodo_Img_Master_3477891Patch.csv filter=lfs diff=lfs merge=lfs -text
 deduplication_process/metadata/Jiggins_Zenodo_Img_Master_3477891Patch_downloaded.csv filter=lfs diff=lfs merge=lfs -text
+#older copies:
+Jiggins_Zenodo_Img_Master.csv filter=lfs diff=lfs merge=lfs -text
+Jiggins_Heliconius_Master.csv filter=lfs diff=lfs merge=lfs -text

Heliconius_img_master.csv

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e61a7a3b58ba98221b2a79e4b8749d5283fe652cf007a01557b795280c99612c
-size 13683618
+oid sha256:a0c3cedf2f8eaeab9763f333a69e157753085a9f3e979b9e49147e895c3427f0
+size 13683556

img_master.csv

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1a82068174884a3663609248aff77c0f292c922582a2a5c7b03b4201489b20fd
-size 16639783
+oid sha256:a0018cb09605ab1eb1ecbfd8dbe631537ea0d3ac6da28afe2e7a2ee897b4113f
+size 16639721

notebooks/Data-gen-1-4.ipynb

@@ -0,0 +1,136 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd\n",
+    "import numpy as np"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Fix \"-\" in \"Collected_by\" Columns in full_master and Heliconius subsets\n",
+    "\n",
+    "See [discussion](https://huggingface.co/datasets/imageomics/Heliconius-Collection_Cambridge-Butterfly/discussions/7). I then tested each column that showed up as \"null\" in type on the dataset viewer (searching for df[col] == \"-\")."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# while main is commit 3264ec40dbc9db025e027ddb2abca0914cd6500f\n",
+    "df = pd.read_csv(\"https://huggingface.co/datasets/imageomics/Heliconius-Collection_Cambridge-Butterfly/resolve/main/img_master.csv\", low_memory= False)\n",
+    "df_heli = pd.read_csv(\"https://huggingface.co/datasets/imageomics/Heliconius-Collection_Cambridge-Butterfly/resolve/main/Heliconius_img_master.csv\", low_memory=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "column Collected_by has 62 instances of '-'\n"
+     ]
+    }
+   ],
+   "source": [
+    "for col in list(df.columns):\n",
+    "    temp = df.loc[df[col] == \"-\"]\n",
+    "    if temp.shape[0] > 0:\n",
+    "        print(f\"column {col} has {temp.shape[0]} instances of '-'\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "column Collected_by has 62 instances of '-'\n"
+     ]
+    }
+   ],
+   "source": [
+    "for col in list(df_heli.columns):\n",
+    "    temp = df_heli.loc[df_heli[col] == \"-\"]\n",
+    "    if temp.shape[0] > 0:\n",
+    "        print(f\"column {col} has {temp.shape[0]} instances of '-'\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Looks like we just need to fix the `Collected_by` column for both CSVs. We'll replace these with `null` so that it doesn't break the dataset viewer."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df.loc[df[\"Collected_by\"] == \"-\", \"Collected_by\"] = np.nan\n",
+    "df_heli.loc[df_heli[\"Collected_by\"] == \"-\", \"Collected_by\"] = np.nan"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Now let's save the fix"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df.to_csv(\"../img_master.csv\", index = False)\n",
+    "df_heli.to_csv(\"../Heliconius_img_master.csv\", index = False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "std",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}