Remove hardcoded API tokens - use environment variables

.gitattributes

@@ -4,5 +4,3 @@
 *.pkl filter=lfs diff=lfs merge=lfs -text
 checkpoint-* filter=lfs diff=lfs merge=lfs -text
 indo-legalbert-with-law-finetuned/* filter=lfs diff=lfs merge=lfs -text
-*.wandb filter=xet -text
-wandb/** filter=xet -text

.gitignore

@@ -1,3 +1,4 @@
 .venv
 __pycache__
-colloquial_datawandb/
+colloquial_data/
+wandb/

11 Indobert-base-p1_without_law.py

@@ -114,13 +114,13 @@ def set_seed(seed: int):
 set_seed(SEED)
 
 # Hugging Face token (optional) - prefer setting this in your environment
-HF_TOKEN = os.environ.get('HF_TOKEN') or "REMOVED_HF_TOKEN"
+HF_TOKEN = os.environ.get('HF_TOKEN')  # Set HF_TOKEN environment variable
 
 # Optional: You can paste a Weights & Biases API key here to avoid interactive prompts.
 # Warning: embedding secrets in source code is not recommended for security reasons.
 # Prefer setting the WANDB_API_KEY environment variable. If you still want to hardcode
 # the key for convenience, paste it below as a string.
-WANDB_API_KEY_INLINE = "REMOVED_WANDB_TOKEN"  # <-- Paste your wandb API key here if you want it embedded (NOT RECOMMENDED)
+WANDB_API_KEY_INLINE = None  # Set WANDB_API_KEY environment variable instead
 
 # W&B project configuration — change if you want a different project/entity
 WANDB_ENTITY = os.environ.get('WANDB_ENTITY') or 'evansteam'
@@ -476,11 +476,12 @@ print(f"   Max length: {MAX_LENGTH}")
 # is present we still proceed without W&B; setting the API key here avoids
 # interactive prompts that can block execution.
 try:
-    # prefer inline key, then env, then fall back to the explicit key provided by user
-    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY") or "REMOVED_WANDB_TOKEN"
+    # prefer inline key, then env
+    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY")
 
     # Export env vars so Trainer/wandb integrations pick them up non-interactively
-    os.environ['WANDB_API_KEY'] = wandb_api
+    if wandb_api:
+        os.environ['WANDB_API_KEY'] = wandb_api
     os.environ['WANDB_ENTITY'] = WANDB_ENTITY
     os.environ['WANDB_PROJECT'] = WANDB_PROJECT
 

12 Indobert-base-p1_with_law.py

@@ -114,13 +114,13 @@ def set_seed(seed: int):
 set_seed(SEED)
 
 # Hugging Face token (optional) - prefer setting this in your environment
-HF_TOKEN = os.environ.get('HF_TOKEN') or "REMOVED_HF_TOKEN"
+HF_TOKEN = os.environ.get('HF_TOKEN')  # Set HF_TOKEN environment variable
 
 # Optional: You can paste a Weights & Biases API key here to avoid interactive prompts.
 # Warning: embedding secrets in source code is not recommended for security reasons.
 # Prefer setting the WANDB_API_KEY environment variable. If you still want to hardcode
 # the key for convenience, paste it below as a string.
-WANDB_API_KEY_INLINE = "REMOVED_WANDB_TOKEN"  # <-- Paste your wandb API key here if you want it embedded (NOT RECOMMENDED)
+WANDB_API_KEY_INLINE = None  # Set WANDB_API_KEY environment variable instead
 
 # W&B project configuration — change if you want a different project/entity
 WANDB_ENTITY = os.environ.get('WANDB_ENTITY') or 'evansteam'
@@ -476,11 +476,12 @@ print(f"   Max length: {MAX_LENGTH}")
 # is present we still proceed without W&B; setting the API key here avoids
 # interactive prompts that can block execution.
 try:
-    # prefer inline key, then env, then fall back to the explicit key provided by user
-    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY") or "REMOVED_WANDB_TOKEN"
+    # prefer inline key, then env
+    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY")
 
     # Export env vars so Trainer/wandb integrations pick them up non-interactively
-    os.environ['WANDB_API_KEY'] = wandb_api
+    if wandb_api:
+        os.environ['WANDB_API_KEY'] = wandb_api
     os.environ['WANDB_ENTITY'] = WANDB_ENTITY
     os.environ['WANDB_PROJECT'] = WANDB_PROJECT
 

21 Indobert-IndoLegalBERT_without_law.py

@@ -114,13 +114,13 @@ def set_seed(seed: int):
 set_seed(SEED)
 
 # Hugging Face token (optional) - prefer setting this in your environment
-HF_TOKEN = os.environ.get('HF_TOKEN') or "REMOVED_HF_TOKEN"
+HF_TOKEN = os.environ.get('HF_TOKEN')  # Set HF_TOKEN environment variable
 
 # Optional: You can paste a Weights & Biases API key here to avoid interactive prompts.
 # Warning: embedding secrets in source code is not recommended for security reasons.
 # Prefer setting the WANDB_API_KEY environment variable. If you still want to hardcode
 # the key for convenience, paste it below as a string.
-WANDB_API_KEY_INLINE = "REMOVED_WANDB_TOKEN"  # <-- Paste your wandb API key here if you want it embedded (NOT RECOMMENDED)
+WANDB_API_KEY_INLINE = None  # Set WANDB_API_KEY environment variable instead
 
 # W&B project configuration — change if you want a different project/entity
 WANDB_ENTITY = os.environ.get('WANDB_ENTITY') or 'evansteam'
@@ -580,11 +580,12 @@ print(f"   Max length: {MAX_LENGTH}")
 # is present we still proceed without W&B; setting the API key here avoids
 # interactive prompts that can block execution.
 try:
-    # prefer inline key, then env, then fall back to the explicit key provided by user
-    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY") or "REMOVED_WANDB_TOKEN"
+    # prefer inline key, then env
+    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY")
 
     # Export env vars so Trainer/wandb integrations pick them up non-interactively
-    os.environ['WANDB_API_KEY'] = wandb_api
+    if wandb_api:
+        os.environ['WANDB_API_KEY'] = wandb_api
     os.environ['WANDB_ENTITY'] = WANDB_ENTITY
     os.environ['WANDB_PROJECT'] = WANDB_PROJECT
 

22 Indobert-IndoLegalBERT_with_law.py

@@ -114,13 +114,13 @@ def set_seed(seed: int):
 set_seed(SEED)
 
 # Hugging Face token (optional) - prefer setting this in your environment
-HF_TOKEN = os.environ.get('HF_TOKEN') or "REMOVED_HF_TOKEN"
+HF_TOKEN = os.environ.get('HF_TOKEN')  # Set HF_TOKEN environment variable
 
 # Optional: You can paste a Weights & Biases API key here to avoid interactive prompts.
 # Warning: embedding secrets in source code is not recommended for security reasons.
 # Prefer setting the WANDB_API_KEY environment variable. If you still want to hardcode
 # the key for convenience, paste it below as a string.
-WANDB_API_KEY_INLINE = "REMOVED_WANDB_TOKEN"  # <-- Paste your wandb API key here if you want it embedded (NOT RECOMMENDED)
+WANDB_API_KEY_INLINE = None  # Set WANDB_API_KEY environment variable instead
 
 # W&B project configuration — change if you want a different project/entity
 WANDB_ENTITY = os.environ.get('WANDB_ENTITY') or 'evansteam'
@@ -580,11 +580,12 @@ print(f"   Max length: {MAX_LENGTH}")
 # is present we still proceed without W&B; setting the API key here avoids
 # interactive prompts that can block execution.
 try:
-    # prefer inline key, then env, then fall back to the explicit key provided by user
-    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY") or "REMOVED_WANDB_TOKEN"
+    # prefer inline key, then env
+    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY")
 
     # Export env vars so Trainer/wandb integrations pick them up non-interactively
-    os.environ['WANDB_API_KEY'] = wandb_api
+    if wandb_api:
+        os.environ['WANDB_API_KEY'] = wandb_api
     os.environ['WANDB_ENTITY'] = WANDB_ENTITY
     os.environ['WANDB_PROJECT'] = WANDB_PROJECT
 

31 Indobert-base-p1_without_law_with_svr.py

@@ -114,13 +114,13 @@ def set_seed(seed: int):
 set_seed(SEED)
 
 # Hugging Face token (optional) - prefer setting this in your environment
-HF_TOKEN = os.environ.get('HF_TOKEN') or "REMOVED_HF_TOKEN"
+HF_TOKEN = os.environ.get('HF_TOKEN')  # Set HF_TOKEN environment variable
 
 # Optional: You can paste a Weights & Biases API key here to avoid interactive prompts.
 # Warning: embedding secrets in source code is not recommended for security reasons.
 # Prefer setting the WANDB_API_KEY environment variable. If you still want to hardcode
 # the key for convenience, paste it below as a string.
-WANDB_API_KEY_INLINE = "REMOVED_WANDB_TOKEN"  # <-- Paste your wandb API key here if you want it embedded (NOT RECOMMENDED)
+WANDB_API_KEY_INLINE = None  # Set WANDB_API_KEY environment variable instead
 
 # W&B project configuration — change if you want a different project/entity
 WANDB_ENTITY = os.environ.get('WANDB_ENTITY') or 'evansteam'
@@ -476,11 +476,12 @@ print(f"   Max length: {MAX_LENGTH}")
 # is present we still proceed without W&B; setting the API key here avoids
 # interactive prompts that can block execution.
 try:
-    # prefer inline key, then env, then fall back to the explicit key provided by user
-    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY") or "REMOVED_WANDB_TOKEN"
+    # prefer inline key, then env
+    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY")
 
     # Export env vars so Trainer/wandb integrations pick them up non-interactively
-    os.environ['WANDB_API_KEY'] = wandb_api
+    if wandb_api:
+        os.environ['WANDB_API_KEY'] = wandb_api
     os.environ['WANDB_ENTITY'] = WANDB_ENTITY
     os.environ['WANDB_PROJECT'] = WANDB_PROJECT
 

32 Indobert-base-p1_with_law_with_svr.py

@@ -114,13 +114,13 @@ def set_seed(seed: int):
 set_seed(SEED)
 
 # Hugging Face token (optional) - prefer setting this in your environment
-HF_TOKEN = os.environ.get('HF_TOKEN') or "REMOVED_HF_TOKEN"
+HF_TOKEN = os.environ.get('HF_TOKEN')  # Set HF_TOKEN environment variable
 
 # Optional: You can paste a Weights & Biases API key here to avoid interactive prompts.
 # Warning: embedding secrets in source code is not recommended for security reasons.
 # Prefer setting the WANDB_API_KEY environment variable. If you still want to hardcode
 # the key for convenience, paste it below as a string.
-WANDB_API_KEY_INLINE = "REMOVED_WANDB_TOKEN"  # <-- Paste your wandb API key here if you want it embedded (NOT RECOMMENDED)
+WANDB_API_KEY_INLINE = None  # Set WANDB_API_KEY environment variable instead
 
 # W&B project configuration — change if you want a different project/entity
 WANDB_ENTITY = os.environ.get('WANDB_ENTITY') or 'evansteam'
@@ -476,11 +476,12 @@ print(f"   Max length: {MAX_LENGTH}")
 # is present we still proceed without W&B; setting the API key here avoids
 # interactive prompts that can block execution.
 try:
-    # prefer inline key, then env, then fall back to the explicit key provided by user
-    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY") or "REMOVED_WANDB_TOKEN"
+    # prefer inline key, then env
+    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY")
 
     # Export env vars so Trainer/wandb integrations pick them up non-interactively
-    os.environ['WANDB_API_KEY'] = wandb_api
+    if wandb_api:
+        os.environ['WANDB_API_KEY'] = wandb_api
     os.environ['WANDB_ENTITY'] = WANDB_ENTITY
     os.environ['WANDB_PROJECT'] = WANDB_PROJECT
 

41 Indobert-base-p1_without_law_mlp.py

@@ -115,13 +115,13 @@ def set_seed(seed: int):
 set_seed(SEED)
 
 # Hugging Face token (optional) - prefer setting this in your environment
-HF_TOKEN = os.environ.get('HF_TOKEN') or "REMOVED_HF_TOKEN"
+HF_TOKEN = os.environ.get('HF_TOKEN')  # Set HF_TOKEN environment variable
 
 # Optional: You can paste a Weights & Biases API key here to avoid interactive prompts.
 # Warning: embedding secrets in source code is not recommended for security reasons.
 # Prefer setting the WANDB_API_KEY environment variable. If you still want to hardcode
 # the key for convenience, paste it below as a string.
-WANDB_API_KEY_INLINE = "REMOVED_WANDB_TOKEN"  # <-- Paste your wandb API key here if you want it embedded (NOT RECOMMENDED)
+WANDB_API_KEY_INLINE = None  # Set WANDB_API_KEY environment variable instead
 
 # W&B project configuration — change if you want a different project/entity
 WANDB_ENTITY = os.environ.get('WANDB_ENTITY') or 'evansteam'
@@ -549,11 +549,12 @@ print(f"   Max length: {MAX_LENGTH}")
 # is present we still proceed without W&B; setting the API key here avoids
 # interactive prompts that can block execution.
 try:
-    # prefer inline key, then env, then fall back to the explicit key provided by user
-    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY") or "REMOVED_WANDB_TOKEN"
+    # prefer inline key, then env
+    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY")
 
     # Export env vars so Trainer/wandb integrations pick them up non-interactively
-    os.environ['WANDB_API_KEY'] = wandb_api
+    if wandb_api:
+        os.environ['WANDB_API_KEY'] = wandb_api
     os.environ['WANDB_ENTITY'] = WANDB_ENTITY
     os.environ['WANDB_PROJECT'] = WANDB_PROJECT
 

42 Indobert-base-p1_with_law_mlp.py

@@ -116,13 +116,13 @@ def set_seed(seed: int):
 set_seed(SEED)
 
 # Hugging Face token (optional) - prefer setting this in your environment
-HF_TOKEN = os.environ.get('HF_TOKEN') or "REMOVED_HF_TOKEN"
+HF_TOKEN = os.environ.get('HF_TOKEN')  # Set HF_TOKEN environment variable
 
 # Optional: You can paste a Weights & Biases API key here to avoid interactive prompts.
 # Warning: embedding secrets in source code is not recommended for security reasons.
 # Prefer setting the WANDB_API_KEY environment variable. If you still want to hardcode
 # the key for convenience, paste it below as a string.
-WANDB_API_KEY_INLINE = "REMOVED_WANDB_TOKEN"  # <-- Paste your wandb API key here if you want it embedded (NOT RECOMMENDED)
+WANDB_API_KEY_INLINE = None  # Set WANDB_API_KEY environment variable instead
 
 # W&B project configuration — change if you want a different project/entity
 WANDB_ENTITY = os.environ.get('WANDB_ENTITY') or 'evansteam'
@@ -550,11 +550,12 @@ print(f"   Max length: {MAX_LENGTH}")
 # is present we still proceed without W&B; setting the API key here avoids
 # interactive prompts that can block execution.
 try:
-    # prefer inline key, then env, then fall back to the explicit key provided by user
-    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY") or "REMOVED_WANDB_TOKEN"
+    # prefer inline key, then env
+    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY")
 
     # Export env vars so Trainer/wandb integrations pick them up non-interactively
-    os.environ['WANDB_API_KEY'] = wandb_api
+    if wandb_api:
+        os.environ['WANDB_API_KEY'] = wandb_api
     os.environ['WANDB_ENTITY'] = WANDB_ENTITY
     os.environ['WANDB_PROJECT'] = WANDB_PROJECT
 

51 Indobert-base-p1_without_law_with_xgboost.py

@@ -48,7 +48,7 @@ def set_seed(seed: int):
 
 set_seed(SEED)
 
-HF_TOKEN = os.environ.get('HF_TOKEN') or "REMOVED_HF_TOKEN"
+HF_TOKEN = os.environ.get('HF_TOKEN')  # Set HF_TOKEN environment variable
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 print(f"Device detected: {device}")
 

52 Indobert-base-p1_with_law_with_xgboost.py

@@ -48,7 +48,7 @@ def set_seed(seed: int):
 
 set_seed(SEED)
 
-HF_TOKEN = os.environ.get('HF_TOKEN') or "REMOVED_HF_TOKEN"
+HF_TOKEN = os.environ.get('HF_TOKEN')  # Set HF_TOKEN environment variable
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 print(f"Device detected: {device}")
 

Indobert-base-p1_with_law_log1p.py

@@ -114,13 +114,13 @@ def set_seed(seed: int):
 set_seed(SEED)
 
 # Hugging Face token (optional) - prefer setting this in your environment
-HF_TOKEN = os.environ.get('HF_TOKEN') or "REMOVED_HF_TOKEN"
+HF_TOKEN = os.environ.get('HF_TOKEN')  # Set HF_TOKEN environment variable
 
 # Optional: You can paste a Weights & Biases API key here to avoid interactive prompts.
 # Warning: embedding secrets in source code is not recommended for security reasons.
 # Prefer setting the WANDB_API_KEY environment variable. If you still want to hardcode
 # the key for convenience, paste it below as a string.
-WANDB_API_KEY_INLINE = "REMOVED_WANDB_TOKEN"  # <-- Paste your wandb API key here if you want it embedded (NOT RECOMMENDED)
+WANDB_API_KEY_INLINE = None  # Set WANDB_API_KEY environment variable instead
 
 # W&B project configuration — change if you want a different project/entity
 WANDB_ENTITY = os.environ.get('WANDB_ENTITY') or 'evansteam'
@@ -507,11 +507,12 @@ print(f"   Max length: {MAX_LENGTH}")
 # is present we still proceed without W&B; setting the API key here avoids
 # interactive prompts that can block execution.
 try:
-    # prefer inline key, then env, then fall back to the explicit key provided by user
-    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY") or "REMOVED_WANDB_TOKEN"
+    # prefer inline key, then env
+    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY")
 
     # Export env vars so Trainer/wandb integrations pick them up non-interactively
-    os.environ['WANDB_API_KEY'] = wandb_api
+    if wandb_api:
+        os.environ['WANDB_API_KEY'] = wandb_api
     os.environ['WANDB_ENTITY'] = WANDB_ENTITY
     os.environ['WANDB_PROJECT'] = WANDB_PROJECT
 

Indobert-base-p1_with_law_nozero.py

@@ -114,13 +114,13 @@ def set_seed(seed: int):
 set_seed(SEED)
 
 # Hugging Face token (optional) - prefer setting this in your environment
-HF_TOKEN = os.environ.get('HF_TOKEN') or "REMOVED_HF_TOKEN"
+HF_TOKEN = os.environ.get('HF_TOKEN')  # Set HF_TOKEN environment variable
 
 # Optional: You can paste a Weights & Biases API key here to avoid interactive prompts.
 # Warning: embedding secrets in source code is not recommended for security reasons.
 # Prefer setting the WANDB_API_KEY environment variable. If you still want to hardcode
 # the key for convenience, paste it below as a string.
-WANDB_API_KEY_INLINE = "REMOVED_WANDB_TOKEN"  # <-- Paste your wandb API key here if you want it embedded (NOT RECOMMENDED)
+WANDB_API_KEY_INLINE = None  # Set WANDB_API_KEY environment variable instead
 
 # W&B project configuration — change if you want a different project/entity
 WANDB_ENTITY = os.environ.get('WANDB_ENTITY') or 'evansteam'
@@ -507,11 +507,12 @@ print(f"   Max length: {MAX_LENGTH}")
 # is present we still proceed without W&B; setting the API key here avoids
 # interactive prompts that can block execution.
 try:
-    # prefer inline key, then env, then fall back to the explicit key provided by user
-    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY") or "REMOVED_WANDB_TOKEN"
+    # prefer inline key, then env
+    wandb_api = WANDB_API_KEY_INLINE or os.environ.get("WANDB_API_KEY")
 
     # Export env vars so Trainer/wandb integrations pick them up non-interactively
-    os.environ['WANDB_API_KEY'] = wandb_api
+    if wandb_api:
+        os.environ['WANDB_API_KEY'] = wandb_api
     os.environ['WANDB_ENTITY'] = WANDB_ENTITY
     os.environ['WANDB_PROJECT'] = WANDB_PROJECT
 

README.md

@@ -10,9 +10,10 @@ curl --proto '=https' --tlsv1.2 -sSf https://raw.githubusercontent.com/huggingfa
 Untuk menyalin repository ini menggunakan token GitHub (cara ini sudah sesuai permintaan):
 
 ```bash
-git clone https://REMOVED_GITHUB_TOKEN@github.com/patrickSevans123/skripsi-regression.git
+git clone https://github.com/patrickSevans123/skripsi-regression.git
 
- git remote set-url origin https://evanslur:REMOVED_HF_TOKEN@huggingface.co/datasets/evanslur/regression-skripsi
+# Set remote with your HF token from environment variable
+git remote set-url origin https://evanslur:$HF_TOKEN@huggingface.co/datasets/evanslur/regression-skripsi
 ```
 
 ---