first commit

.gitignore

@@ -0,0 +1,3 @@
+*.DS_Store
+secrets.toml
+create_sample_data.py

Dockerfile

@@ -1,4 +1,4 @@
-FROM python:3.9-slim
+FROM python:3.12-slim
 
 WORKDIR /app
 

README.md

@@ -1,5 +1,5 @@
 ---
-title: UPTranslate
+title: UP Translate
 emoji: 🚀
 colorFrom: red
 colorTo: red

requirements.txt

@@ -1,4 +1,70 @@
-altair
-pandas
-streamlit
-pandas
+altair==5.5.0
+annotated-types==0.7.0
+anyio==4.10.0
+attrs==25.3.0
+blinker==1.9.0
+cachetools==5.5.2
+certifi==2025.8.3
+charset-normalizer==3.4.3
+click==8.2.1
+contourpy==1.3.3
+cycler==0.12.1
+distro==1.9.0
+filelock==3.18.0
+fonttools==4.59.0
+fsspec==2025.7.0
+gitdb==4.0.12
+GitPython==3.1.45
+google-auth==2.40.3
+google-genai==1.29.0
+h11==0.16.0
+hf-xet==1.1.7
+httpcore==1.0.9
+httpx==0.28.1
+huggingface-hub==0.34.4
+idna==3.10
+Jinja2==3.1.6
+jiter==0.10.0
+jsonschema==4.25.0
+jsonschema-specifications==2025.4.1
+kiwisolver==1.4.9
+MarkupSafe==3.0.2
+matplotlib==3.10.5
+narwhals==2.1.0
+numpy==2.3.2
+openai==1.99.8
+packaging==25.0
+pandas==2.3.1
+pillow==11.3.0
+plotly==6.2.0
+protobuf==6.31.1
+pyarrow==21.0.0
+pyasn1==0.6.1
+pyasn1_modules==0.4.2
+pydantic==2.11.7
+pydantic_core==2.33.2
+pydeck==0.9.1
+pyparsing==3.2.3
+python-dateutil==2.9.0.post0
+pytz==2025.2
+PyYAML==6.0.2
+referencing==0.36.2
+requests==2.32.4
+rpds-py==0.27.0
+rsa==4.9.1
+setuptools==78.1.1
+six==1.17.0
+smmap==5.0.2
+sniffio==1.3.1
+streamlit==1.48.0
+tenacity==9.1.2
+toml==0.10.2
+tornado==6.5.2
+tqdm==4.67.1
+typing-inspection==0.4.1
+typing_extensions==4.14.1
+tzdata==2025.2
+urllib3==2.5.0
+websockets==15.0.1
+wheel==0.45.1
+wordcloud==1.9.4

src/.streamlit/config.toml

@@ -0,0 +1,14 @@
+[server]
+headless = true
+port = 8501
+enableCORS = false
+enableXsrfProtection = false
+
+[browser]
+gatherUsageStats = false
+
+[theme]
+primaryColor = "#3b82f6"
+backgroundColor = "#ffffff"
+secondaryBackgroundColor = "#f9fafb"
+textColor = "#1f2937"

src/data/bleu_scores.csv

@@ -0,0 +1,109 @@
+comparison_pair,bleu_score,language,domain,num_sentences
+afr_human_vs_afr_revised,0.8011,Afrikaans,Overall,96
+afr_human_vs_afr_revised,0.8176,Afrikaans,Economics,52
+afr_human_vs_afr_revised,0.7445,Afrikaans,Computer Science,26
+afr_human_vs_afr_revised,0.8353,Afrikaans,Electrical Electronic and Computer Engineering,6
+afr_human_vs_afr_gemini,0.6156,Afrikaans,Overall,306
+afr_human_vs_afr_gemini,0.7456,Afrikaans,Historical and Heritage Studies,19
+afr_human_vs_afr_gemini,0.6069,Afrikaans,Economics,132
+afr_human_vs_afr_gemini,0.7174,Afrikaans,Information Science,13
+afr_human_vs_afr_gemini,0.664,Afrikaans,Informatics,12
+afr_human_vs_afr_gemini,0.6741,Afrikaans,Nursing Science,11
+afr_human_vs_afr_gemini,0.5194,Afrikaans,Computer Science,46
+afr_human_vs_afr_gemini,0.5548,Afrikaans,"Geography, Geoinformatics and Meteorology",11
+afr_human_vs_afr_gemini,0.6021,Afrikaans,Electrical Electronic and Computer Engineering,41
+afr_human_vs_afr_gemini,0.8021,Afrikaans,Microbiology and Plant Pathology,13
+afr_human_vs_afr_gemini,0.5323,Afrikaans,Statistics,8
+afr_human_vs_afr_gpt,0.6174,Afrikaans,Overall,306
+afr_human_vs_afr_gpt,0.7584,Afrikaans,Historical and Heritage Studies,19
+afr_human_vs_afr_gpt,0.5969,Afrikaans,Economics,132
+afr_human_vs_afr_gpt,0.804,Afrikaans,Information Science,13
+afr_human_vs_afr_gpt,0.674,Afrikaans,Informatics,12
+afr_human_vs_afr_gpt,0.7185,Afrikaans,Nursing Science,11
+afr_human_vs_afr_gpt,0.5063,Afrikaans,Computer Science,46
+afr_human_vs_afr_gpt,0.6468,Afrikaans,"Geography, Geoinformatics and Meteorology",11
+afr_human_vs_afr_gpt,0.5949,Afrikaans,Electrical Electronic and Computer Engineering,41
+afr_human_vs_afr_gpt,0.7844,Afrikaans,Microbiology and Plant Pathology,13
+afr_human_vs_afr_gpt,0.5348,Afrikaans,Statistics,8
+nso_human_vs_nso_revised,0.3825,Northern Sotho,Overall,55
+nso_human_vs_nso_revised,0.4432,Northern Sotho,Historical and Heritage Studies,18
+nso_human_vs_nso_revised,0.461,Northern Sotho,Economics,10
+nso_human_vs_nso_revised,0.1491,Northern Sotho,Computer Science,8
+nso_human_vs_nso_revised,0.5369,Northern Sotho,Statistics,6
+nso_human_vs_nso_gemini,0.3031,Northern Sotho,Overall,307
+nso_human_vs_nso_gemini,0.2998,Northern Sotho,Historical and Heritage Studies,19
+nso_human_vs_nso_gemini,0.3178,Northern Sotho,Economics,133
+nso_human_vs_nso_gemini,0.2806,Northern Sotho,Information Science,13
+nso_human_vs_nso_gemini,0.4086,Northern Sotho,Informatics,12
+nso_human_vs_nso_gemini,0.2898,Northern Sotho,Nursing Science,11
+nso_human_vs_nso_gemini,0.2825,Northern Sotho,Computer Science,46
+nso_human_vs_nso_gemini,0.2595,Northern Sotho,"Geography, Geoinformatics and Meteorology",11
+nso_human_vs_nso_gemini,0.2148,Northern Sotho,Electrical Electronic and Computer Engineering,41
+nso_human_vs_nso_gemini,0.4771,Northern Sotho,Microbiology and Plant Pathology,13
+nso_human_vs_nso_gemini,0.314,Northern Sotho,Statistics,8
+nso_human_vs_nso_gpt,0.2526,Northern Sotho,Overall,307
+nso_human_vs_nso_gpt,0.3058,Northern Sotho,Historical and Heritage Studies,19
+nso_human_vs_nso_gpt,0.2753,Northern Sotho,Economics,133
+nso_human_vs_nso_gpt,0.2502,Northern Sotho,Information Science,13
+nso_human_vs_nso_gpt,0.2465,Northern Sotho,Informatics,12
+nso_human_vs_nso_gpt,0.2596,Northern Sotho,Nursing Science,11
+nso_human_vs_nso_gpt,0.1992,Northern Sotho,Computer Science,46
+nso_human_vs_nso_gpt,0.2172,Northern Sotho,"Geography, Geoinformatics and Meteorology",11
+nso_human_vs_nso_gpt,0.2038,Northern Sotho,Electrical Electronic and Computer Engineering,41
+nso_human_vs_nso_gpt,0.3609,Northern Sotho,Microbiology and Plant Pathology,13
+nso_human_vs_nso_gpt,0.1826,Northern Sotho,Statistics,8
+nso_human_vs_nso_nllb,0.266,Northern Sotho,Overall,307
+nso_human_vs_nso_nllb,0.2611,Northern Sotho,Historical and Heritage Studies,19
+nso_human_vs_nso_nllb,0.2938,Northern Sotho,Economics,133
+nso_human_vs_nso_nllb,0.2718,Northern Sotho,Information Science,13
+nso_human_vs_nso_nllb,0.2987,Northern Sotho,Informatics,12
+nso_human_vs_nso_nllb,0.2908,Northern Sotho,Nursing Science,11
+nso_human_vs_nso_nllb,0.2119,Northern Sotho,Computer Science,46
+nso_human_vs_nso_nllb,0.221,Northern Sotho,"Geography, Geoinformatics and Meteorology",11
+nso_human_vs_nso_nllb,0.2069,Northern Sotho,Electrical Electronic and Computer Engineering,41
+nso_human_vs_nso_nllb,0.3858,Northern Sotho,Microbiology and Plant Pathology,13
+nso_human_vs_nso_nllb,0.202,Northern Sotho,Statistics,8
+zul_human_vs_zul_revised,0.3816,isiZulu,Overall,285
+zul_human_vs_zul_revised,0.4277,isiZulu,Historical and Heritage Studies,18
+zul_human_vs_zul_revised,0.3973,isiZulu,Economics,128
+zul_human_vs_zul_revised,0.5029,isiZulu,Information Science,11
+zul_human_vs_zul_revised,0.4266,isiZulu,Informatics,10
+zul_human_vs_zul_revised,0.5591,isiZulu,Nursing Science,7
+zul_human_vs_zul_revised,0.1769,isiZulu,Computer Science,45
+zul_human_vs_zul_revised,0.3338,isiZulu,"Geography, Geoinformatics and Meteorology",10
+zul_human_vs_zul_revised,0.4241,isiZulu,Electrical Electronic and Computer Engineering,36
+zul_human_vs_zul_revised,0.5684,isiZulu,Microbiology and Plant Pathology,12
+zul_human_vs_zul_revised,0.3879,isiZulu,Statistics,8
+zul_human_vs_zul_gemini,0.0907,isiZulu,Overall,307
+zul_human_vs_zul_gemini,0.0972,isiZulu,Historical and Heritage Studies,19
+zul_human_vs_zul_gemini,0.0832,isiZulu,Economics,133
+zul_human_vs_zul_gemini,0.1764,isiZulu,Information Science,13
+zul_human_vs_zul_gemini,0.0821,isiZulu,Informatics,12
+zul_human_vs_zul_gemini,0.1561,isiZulu,Nursing Science,11
+zul_human_vs_zul_gemini,0.0692,isiZulu,Computer Science,46
+zul_human_vs_zul_gemini,0.0819,isiZulu,"Geography, Geoinformatics and Meteorology",11
+zul_human_vs_zul_gemini,0.0809,isiZulu,Electrical Electronic and Computer Engineering,41
+zul_human_vs_zul_gemini,0.177,isiZulu,Microbiology and Plant Pathology,13
+zul_human_vs_zul_gemini,0.0313,isiZulu,Statistics,8
+zul_human_vs_zul_gpt,0.0856,isiZulu,Overall,307
+zul_human_vs_zul_gpt,0.1059,isiZulu,Historical and Heritage Studies,19
+zul_human_vs_zul_gpt,0.078,isiZulu,Economics,133
+zul_human_vs_zul_gpt,0.1177,isiZulu,Information Science,13
+zul_human_vs_zul_gpt,0.112,isiZulu,Informatics,12
+zul_human_vs_zul_gpt,0.1178,isiZulu,Nursing Science,11
+zul_human_vs_zul_gpt,0.0752,isiZulu,Computer Science,46
+zul_human_vs_zul_gpt,0.1032,isiZulu,"Geography, Geoinformatics and Meteorology",11
+zul_human_vs_zul_gpt,0.0877,isiZulu,Electrical Electronic and Computer Engineering,41
+zul_human_vs_zul_gpt,0.096,isiZulu,Microbiology and Plant Pathology,13
+zul_human_vs_zul_gpt,0.0355,isiZulu,Statistics,8
+zul_human_vs_zul_nllb,0.0909,isiZulu,Overall,307
+zul_human_vs_zul_nllb,0.0877,isiZulu,Historical and Heritage Studies,19
+zul_human_vs_zul_nllb,0.0855,isiZulu,Economics,133
+zul_human_vs_zul_nllb,0.1456,isiZulu,Information Science,13
+zul_human_vs_zul_nllb,0.1207,isiZulu,Informatics,12
+zul_human_vs_zul_nllb,0.1548,isiZulu,Nursing Science,11
+zul_human_vs_zul_nllb,0.0678,isiZulu,Computer Science,46
+zul_human_vs_zul_nllb,0.1022,isiZulu,"Geography, Geoinformatics and Meteorology",11
+zul_human_vs_zul_nllb,0.0873,isiZulu,Electrical Electronic and Computer Engineering,41
+zul_human_vs_zul_nllb,0.1337,isiZulu,Microbiology and Plant Pathology,13
+zul_human_vs_zul_nllb,0.0333,isiZulu,Statistics,8

src/data/chrf_scores.csv

@@ -0,0 +1,119 @@
+comparison_pair,chrf_score,language,domain,num_sentences
+afr_human_vs_afr_revised,0.9262,Afrikaans,Overall,96
+afr_human_vs_afr_revised,0.9393,Afrikaans,Historical and Heritage Studies,3
+afr_human_vs_afr_revised,0.9285,Afrikaans,Economics,52
+afr_human_vs_afr_revised,0.9008,Afrikaans,Information Science,2
+afr_human_vs_afr_revised,0.9509,Afrikaans,Informatics,2
+afr_human_vs_afr_revised,0.9188,Afrikaans,Computer Science,26
+afr_human_vs_afr_revised,0.9327,Afrikaans,"Geography, Geoinformatics and Meteorology",2
+afr_human_vs_afr_revised,0.9283,Afrikaans,Electrical Electronic and Computer Engineering,6
+afr_human_vs_afr_revised,0.9311,Afrikaans,Microbiology and Plant Pathology,3
+afr_human_vs_afr_gemini,0.8473,Afrikaans,Overall,306
+afr_human_vs_afr_gemini,0.8963,Afrikaans,Historical and Heritage Studies,19
+afr_human_vs_afr_gemini,0.8402,Afrikaans,Economics,132
+afr_human_vs_afr_gemini,0.9234,Afrikaans,Information Science,13
+afr_human_vs_afr_gemini,0.8674,Afrikaans,Informatics,12
+afr_human_vs_afr_gemini,0.861,Afrikaans,Nursing Science,11
+afr_human_vs_afr_gemini,0.8062,Afrikaans,Computer Science,46
+afr_human_vs_afr_gemini,0.8123,Afrikaans,"Geography, Geoinformatics and Meteorology",11
+afr_human_vs_afr_gemini,0.8644,Afrikaans,Electrical Electronic and Computer Engineering,41
+afr_human_vs_afr_gemini,0.8904,Afrikaans,Microbiology and Plant Pathology,13
+afr_human_vs_afr_gemini,0.803,Afrikaans,Statistics,8
+afr_human_vs_afr_gpt,0.8415,Afrikaans,Overall,306
+afr_human_vs_afr_gpt,0.907,Afrikaans,Historical and Heritage Studies,19
+afr_human_vs_afr_gpt,0.8258,Afrikaans,Economics,132
+afr_human_vs_afr_gpt,0.9429,Afrikaans,Information Science,13
+afr_human_vs_afr_gpt,0.8827,Afrikaans,Informatics,12
+afr_human_vs_afr_gpt,0.8912,Afrikaans,Nursing Science,11
+afr_human_vs_afr_gpt,0.7763,Afrikaans,Computer Science,46
+afr_human_vs_afr_gpt,0.8542,Afrikaans,"Geography, Geoinformatics and Meteorology",11
+afr_human_vs_afr_gpt,0.8601,Afrikaans,Electrical Electronic and Computer Engineering,41
+afr_human_vs_afr_gpt,0.9066,Afrikaans,Microbiology and Plant Pathology,13
+afr_human_vs_afr_gpt,0.8076,Afrikaans,Statistics,8
+nso_human_vs_nso_revised,0.5745,Northern Sotho,Overall,55
+nso_human_vs_nso_revised,0.6954,Northern Sotho,Historical and Heritage Studies,18
+nso_human_vs_nso_revised,0.6416,Northern Sotho,Economics,10
+nso_human_vs_nso_revised,0.2633,Northern Sotho,Information Science,4
+nso_human_vs_nso_revised,0.7982,Northern Sotho,Informatics,2
+nso_human_vs_nso_revised,0.5552,Northern Sotho,Nursing Science,3
+nso_human_vs_nso_revised,0.344,Northern Sotho,Computer Science,8
+nso_human_vs_nso_revised,0.3796,Northern Sotho,"Geography, Geoinformatics and Meteorology",2
+nso_human_vs_nso_revised,0.3478,Northern Sotho,Microbiology and Plant Pathology,2
+nso_human_vs_nso_revised,0.69,Northern Sotho,Statistics,6
+nso_human_vs_nso_gemini,0.5896,Northern Sotho,Overall,307
+nso_human_vs_nso_gemini,0.5832,Northern Sotho,Historical and Heritage Studies,19
+nso_human_vs_nso_gemini,0.6014,Northern Sotho,Economics,133
+nso_human_vs_nso_gemini,0.5658,Northern Sotho,Information Science,13
+nso_human_vs_nso_gemini,0.676,Northern Sotho,Informatics,12
+nso_human_vs_nso_gemini,0.5761,Northern Sotho,Nursing Science,11
+nso_human_vs_nso_gemini,0.5671,Northern Sotho,Computer Science,46
+nso_human_vs_nso_gemini,0.5634,Northern Sotho,"Geography, Geoinformatics and Meteorology",11
+nso_human_vs_nso_gemini,0.5327,Northern Sotho,Electrical Electronic and Computer Engineering,41
+nso_human_vs_nso_gemini,0.6915,Northern Sotho,Microbiology and Plant Pathology,13
+nso_human_vs_nso_gemini,0.6283,Northern Sotho,Statistics,8
+nso_human_vs_nso_gpt,0.5434,Northern Sotho,Overall,307
+nso_human_vs_nso_gpt,0.5656,Northern Sotho,Historical and Heritage Studies,19
+nso_human_vs_nso_gpt,0.5761,Northern Sotho,Economics,133
+nso_human_vs_nso_gpt,0.5459,Northern Sotho,Information Science,13
+nso_human_vs_nso_gpt,0.5263,Northern Sotho,Informatics,12
+nso_human_vs_nso_gpt,0.5661,Northern Sotho,Nursing Science,11
+nso_human_vs_nso_gpt,0.4661,Northern Sotho,Computer Science,46
+nso_human_vs_nso_gpt,0.526,Northern Sotho,"Geography, Geoinformatics and Meteorology",11
+nso_human_vs_nso_gpt,0.499,Northern Sotho,Electrical Electronic and Computer Engineering,41
+nso_human_vs_nso_gpt,0.6324,Northern Sotho,Microbiology and Plant Pathology,13
+nso_human_vs_nso_gpt,0.4877,Northern Sotho,Statistics,8
+nso_human_vs_nso_nllb,0.5688,Northern Sotho,Overall,307
+nso_human_vs_nso_nllb,0.5786,Northern Sotho,Historical and Heritage Studies,19
+nso_human_vs_nso_nllb,0.588,Northern Sotho,Economics,133
+nso_human_vs_nso_nllb,0.5736,Northern Sotho,Information Science,13
+nso_human_vs_nso_nllb,0.6325,Northern Sotho,Informatics,12
+nso_human_vs_nso_nllb,0.6527,Northern Sotho,Nursing Science,11
+nso_human_vs_nso_nllb,0.5006,Northern Sotho,Computer Science,46
+nso_human_vs_nso_nllb,0.5273,Northern Sotho,"Geography, Geoinformatics and Meteorology",11
+nso_human_vs_nso_nllb,0.5211,Northern Sotho,Electrical Electronic and Computer Engineering,41
+nso_human_vs_nso_nllb,0.6732,Northern Sotho,Microbiology and Plant Pathology,13
+nso_human_vs_nso_nllb,0.5323,Northern Sotho,Statistics,8
+zul_human_vs_zul_revised,0.7141,isiZulu,Overall,285
+zul_human_vs_zul_revised,0.7486,isiZulu,Historical and Heritage Studies,18
+zul_human_vs_zul_revised,0.7612,isiZulu,Economics,128
+zul_human_vs_zul_revised,0.7933,isiZulu,Information Science,11
+zul_human_vs_zul_revised,0.7246,isiZulu,Informatics,10
+zul_human_vs_zul_revised,0.8462,isiZulu,Nursing Science,7
+zul_human_vs_zul_revised,0.5057,isiZulu,Computer Science,45
+zul_human_vs_zul_revised,0.6284,isiZulu,"Geography, Geoinformatics and Meteorology",10
+zul_human_vs_zul_revised,0.7279,isiZulu,Electrical Electronic and Computer Engineering,36
+zul_human_vs_zul_revised,0.7971,isiZulu,Microbiology and Plant Pathology,12
+zul_human_vs_zul_revised,0.7401,isiZulu,Statistics,8
+zul_human_vs_zul_gemini,0.4992,isiZulu,Overall,307
+zul_human_vs_zul_gemini,0.529,isiZulu,Historical and Heritage Studies,19
+zul_human_vs_zul_gemini,0.5243,isiZulu,Economics,133
+zul_human_vs_zul_gemini,0.5735,isiZulu,Information Science,13
+zul_human_vs_zul_gemini,0.5031,isiZulu,Informatics,12
+zul_human_vs_zul_gemini,0.6054,isiZulu,Nursing Science,11
+zul_human_vs_zul_gemini,0.4244,isiZulu,Computer Science,46
+zul_human_vs_zul_gemini,0.3796,isiZulu,"Geography, Geoinformatics and Meteorology",11
+zul_human_vs_zul_gemini,0.4656,isiZulu,Electrical Electronic and Computer Engineering,41
+zul_human_vs_zul_gemini,0.5445,isiZulu,Microbiology and Plant Pathology,13
+zul_human_vs_zul_gemini,0.4321,isiZulu,Statistics,8
+zul_human_vs_zul_gpt,0.4835,isiZulu,Overall,307
+zul_human_vs_zul_gpt,0.5149,isiZulu,Historical and Heritage Studies,19
+zul_human_vs_zul_gpt,0.5069,isiZulu,Economics,133
+zul_human_vs_zul_gpt,0.5513,isiZulu,Information Science,13
+zul_human_vs_zul_gpt,0.5362,isiZulu,Informatics,12
+zul_human_vs_zul_gpt,0.5969,isiZulu,Nursing Science,11
+zul_human_vs_zul_gpt,0.4024,isiZulu,Computer Science,46
+zul_human_vs_zul_gpt,0.3585,isiZulu,"Geography, Geoinformatics and Meteorology",11
+zul_human_vs_zul_gpt,0.4605,isiZulu,Electrical Electronic and Computer Engineering,41
+zul_human_vs_zul_gpt,0.4868,isiZulu,Microbiology and Plant Pathology,13
+zul_human_vs_zul_gpt,0.4248,isiZulu,Statistics,8
+zul_human_vs_zul_nllb,0.4881,isiZulu,Overall,307
+zul_human_vs_zul_nllb,0.514,isiZulu,Historical and Heritage Studies,19
+zul_human_vs_zul_nllb,0.515,isiZulu,Economics,133
+zul_human_vs_zul_nllb,0.5514,isiZulu,Information Science,13
+zul_human_vs_zul_nllb,0.5189,isiZulu,Informatics,12
+zul_human_vs_zul_nllb,0.6127,isiZulu,Nursing Science,11
+zul_human_vs_zul_nllb,0.4027,isiZulu,Computer Science,46
+zul_human_vs_zul_nllb,0.3622,isiZulu,"Geography, Geoinformatics and Meteorology",11
+zul_human_vs_zul_nllb,0.4655,isiZulu,Electrical Electronic and Computer Engineering,41
+zul_human_vs_zul_nllb,0.4983,isiZulu,Microbiology and Plant Pathology,13
+zul_human_vs_zul_nllb,0.4202,isiZulu,Statistics,8

src/data/comet_scores.csv

@@ -0,0 +1,119 @@
+comparison_pair,comet_score,language,domain,num_sentences
+afr_human_vs_afr_revised,0.7352,Afrikaans,Overall,96
+afr_human_vs_afr_revised,0.7537,Afrikaans,Historical and Heritage Studies,3
+afr_human_vs_afr_revised,0.7432,Afrikaans,Economics,52
+afr_human_vs_afr_revised,0.8113,Afrikaans,Information Science,2
+afr_human_vs_afr_revised,0.7563,Afrikaans,Informatics,2
+afr_human_vs_afr_revised,0.714,Afrikaans,Computer Science,26
+afr_human_vs_afr_revised,0.6041,Afrikaans,"Geography, Geoinformatics and Meteorology",2
+afr_human_vs_afr_revised,0.761,Afrikaans,Electrical Electronic and Computer Engineering,6
+afr_human_vs_afr_revised,0.7336,Afrikaans,Microbiology and Plant Pathology,3
+afr_human_vs_afr_gemini,0.7557,Afrikaans,Overall,306
+afr_human_vs_afr_gemini,0.7848,Afrikaans,Historical and Heritage Studies,19
+afr_human_vs_afr_gemini,0.7579,Afrikaans,Economics,132
+afr_human_vs_afr_gemini,0.7995,Afrikaans,Information Science,13
+afr_human_vs_afr_gemini,0.8137,Afrikaans,Informatics,12
+afr_human_vs_afr_gemini,0.7744,Afrikaans,Nursing Science,11
+afr_human_vs_afr_gemini,0.7334,Afrikaans,Computer Science,46
+afr_human_vs_afr_gemini,0.6529,Afrikaans,"Geography, Geoinformatics and Meteorology",11
+afr_human_vs_afr_gemini,0.7551,Afrikaans,Electrical Electronic and Computer Engineering,41
+afr_human_vs_afr_gemini,0.8092,Afrikaans,Microbiology and Plant Pathology,13
+afr_human_vs_afr_gemini,0.6535,Afrikaans,Statistics,8
+afr_human_vs_afr_gpt,0.7536,Afrikaans,Overall,306
+afr_human_vs_afr_gpt,0.7778,Afrikaans,Historical and Heritage Studies,19
+afr_human_vs_afr_gpt,0.76,Afrikaans,Economics,132
+afr_human_vs_afr_gpt,0.8068,Afrikaans,Information Science,13
+afr_human_vs_afr_gpt,0.8077,Afrikaans,Informatics,12
+afr_human_vs_afr_gpt,0.7733,Afrikaans,Nursing Science,11
+afr_human_vs_afr_gpt,0.7259,Afrikaans,Computer Science,46
+afr_human_vs_afr_gpt,0.6414,Afrikaans,"Geography, Geoinformatics and Meteorology",11
+afr_human_vs_afr_gpt,0.7513,Afrikaans,Electrical Electronic and Computer Engineering,41
+afr_human_vs_afr_gpt,0.8053,Afrikaans,Microbiology and Plant Pathology,13
+afr_human_vs_afr_gpt,0.6374,Afrikaans,Statistics,8
+nso_human_vs_nso_revised,0.5325,Northern Sotho,Overall,55
+nso_human_vs_nso_revised,0.6267,Northern Sotho,Historical and Heritage Studies,18
+nso_human_vs_nso_revised,0.6125,Northern Sotho,Economics,10
+nso_human_vs_nso_revised,0.322,Northern Sotho,Information Science,4
+nso_human_vs_nso_revised,0.6908,Northern Sotho,Informatics,2
+nso_human_vs_nso_revised,0.4732,Northern Sotho,Nursing Science,3
+nso_human_vs_nso_revised,0.457,Northern Sotho,Computer Science,8
+nso_human_vs_nso_revised,0.4298,Northern Sotho,"Geography, Geoinformatics and Meteorology",2
+nso_human_vs_nso_revised,0.3199,Northern Sotho,Microbiology and Plant Pathology,2
+nso_human_vs_nso_revised,0.4392,Northern Sotho,Statistics,6
+nso_human_vs_nso_gemini,0.6576,Northern Sotho,Overall,307
+nso_human_vs_nso_gemini,0.6815,Northern Sotho,Historical and Heritage Studies,19
+nso_human_vs_nso_gemini,0.6678,Northern Sotho,Economics,133
+nso_human_vs_nso_gemini,0.7196,Northern Sotho,Information Science,13
+nso_human_vs_nso_gemini,0.7279,Northern Sotho,Informatics,12
+nso_human_vs_nso_gemini,0.6901,Northern Sotho,Nursing Science,11
+nso_human_vs_nso_gemini,0.6354,Northern Sotho,Computer Science,46
+nso_human_vs_nso_gemini,0.5456,Northern Sotho,"Geography, Geoinformatics and Meteorology",11
+nso_human_vs_nso_gemini,0.6331,Northern Sotho,Electrical Electronic and Computer Engineering,41
+nso_human_vs_nso_gemini,0.7121,Northern Sotho,Microbiology and Plant Pathology,13
+nso_human_vs_nso_gemini,0.4975,Northern Sotho,Statistics,8
+nso_human_vs_nso_gpt,0.6461,Northern Sotho,Overall,307
+nso_human_vs_nso_gpt,0.6941,Northern Sotho,Historical and Heritage Studies,19
+nso_human_vs_nso_gpt,0.6593,Northern Sotho,Economics,133
+nso_human_vs_nso_gpt,0.6923,Northern Sotho,Information Science,13
+nso_human_vs_nso_gpt,0.7227,Northern Sotho,Informatics,12
+nso_human_vs_nso_gpt,0.6724,Northern Sotho,Nursing Science,11
+nso_human_vs_nso_gpt,0.6159,Northern Sotho,Computer Science,46
+nso_human_vs_nso_gpt,0.5706,Northern Sotho,"Geography, Geoinformatics and Meteorology",11
+nso_human_vs_nso_gpt,0.6115,Northern Sotho,Electrical Electronic and Computer Engineering,41
+nso_human_vs_nso_gpt,0.6755,Northern Sotho,Microbiology and Plant Pathology,13
+nso_human_vs_nso_gpt,0.494,Northern Sotho,Statistics,8
+nso_human_vs_nso_nllb,0.6234,Northern Sotho,Overall,307
+nso_human_vs_nso_nllb,0.6435,Northern Sotho,Historical and Heritage Studies,19
+nso_human_vs_nso_nllb,0.6413,Northern Sotho,Economics,133
+nso_human_vs_nso_nllb,0.6693,Northern Sotho,Information Science,13
+nso_human_vs_nso_nllb,0.7008,Northern Sotho,Informatics,12
+nso_human_vs_nso_nllb,0.6154,Northern Sotho,Nursing Science,11
+nso_human_vs_nso_nllb,0.5793,Northern Sotho,Computer Science,46
+nso_human_vs_nso_nllb,0.5199,Northern Sotho,"Geography, Geoinformatics and Meteorology",11
+nso_human_vs_nso_nllb,0.5986,Northern Sotho,Electrical Electronic and Computer Engineering,41
+nso_human_vs_nso_nllb,0.6945,Northern Sotho,Microbiology and Plant Pathology,13
+nso_human_vs_nso_nllb,0.5055,Northern Sotho,Statistics,8
+zul_human_vs_zul_revised,0.5969,isiZulu,Overall,285
+zul_human_vs_zul_revised,0.6104,isiZulu,Historical and Heritage Studies,18
+zul_human_vs_zul_revised,0.6098,isiZulu,Economics,128
+zul_human_vs_zul_revised,0.6708,isiZulu,Information Science,11
+zul_human_vs_zul_revised,0.6003,isiZulu,Informatics,10
+zul_human_vs_zul_revised,0.6473,isiZulu,Nursing Science,7
+zul_human_vs_zul_revised,0.5703,isiZulu,Computer Science,45
+zul_human_vs_zul_revised,0.5336,isiZulu,"Geography, Geoinformatics and Meteorology",10
+zul_human_vs_zul_revised,0.5658,isiZulu,Electrical Electronic and Computer Engineering,36
+zul_human_vs_zul_revised,0.6909,isiZulu,Microbiology and Plant Pathology,12
+zul_human_vs_zul_revised,0.4388,isiZulu,Statistics,8
+zul_human_vs_zul_gemini,0.6711,isiZulu,Overall,307
+zul_human_vs_zul_gemini,0.6935,isiZulu,Historical and Heritage Studies,19
+zul_human_vs_zul_gemini,0.6927,isiZulu,Economics,133
+zul_human_vs_zul_gemini,0.6992,isiZulu,Information Science,13
+zul_human_vs_zul_gemini,0.7448,isiZulu,Informatics,12
+zul_human_vs_zul_gemini,0.73,isiZulu,Nursing Science,11
+zul_human_vs_zul_gemini,0.628,isiZulu,Computer Science,46
+zul_human_vs_zul_gemini,0.5512,isiZulu,"Geography, Geoinformatics and Meteorology",11
+zul_human_vs_zul_gemini,0.6392,isiZulu,Electrical Electronic and Computer Engineering,41
+zul_human_vs_zul_gemini,0.7165,isiZulu,Microbiology and Plant Pathology,13
+zul_human_vs_zul_gemini,0.522,isiZulu,Statistics,8
+zul_human_vs_zul_gpt,0.6573,isiZulu,Overall,307
+zul_human_vs_zul_gpt,0.6887,isiZulu,Historical and Heritage Studies,19
+zul_human_vs_zul_gpt,0.672,isiZulu,Economics,133
+zul_human_vs_zul_gpt,0.7049,isiZulu,Information Science,13
+zul_human_vs_zul_gpt,0.7064,isiZulu,Informatics,12
+zul_human_vs_zul_gpt,0.7019,isiZulu,Nursing Science,11
+zul_human_vs_zul_gpt,0.6348,isiZulu,Computer Science,46
+zul_human_vs_zul_gpt,0.5503,isiZulu,"Geography, Geoinformatics and Meteorology",11
+zul_human_vs_zul_gpt,0.6167,isiZulu,Electrical Electronic and Computer Engineering,41
+zul_human_vs_zul_gpt,0.702,isiZulu,Microbiology and Plant Pathology,13
+zul_human_vs_zul_gpt,0.5381,isiZulu,Statistics,8
+zul_human_vs_zul_nllb,0.6326,isiZulu,Overall,307
+zul_human_vs_zul_nllb,0.6477,isiZulu,Historical and Heritage Studies,19
+zul_human_vs_zul_nllb,0.6441,isiZulu,Economics,133
+zul_human_vs_zul_nllb,0.7024,isiZulu,Information Science,13
+zul_human_vs_zul_nllb,0.6976,isiZulu,Informatics,12
+zul_human_vs_zul_nllb,0.7014,isiZulu,Nursing Science,11
+zul_human_vs_zul_nllb,0.6036,isiZulu,Computer Science,46
+zul_human_vs_zul_nllb,0.5395,isiZulu,"Geography, Geoinformatics and Meteorology",11
+zul_human_vs_zul_nllb,0.5939,isiZulu,Electrical Electronic and Computer Engineering,41
+zul_human_vs_zul_nllb,0.6935,isiZulu,Microbiology and Plant Pathology,13
+zul_human_vs_zul_nllb,0.4938,isiZulu,Statistics,8