Upload 4 files

Upload additional requirements

config.json

@@ -0,0 +1,76 @@
+{
+    "format": "piisa:config:full:v1",
+    "config": [
+        {
+            "format": "piisa:config:pii-extract-plg-transformers:main:v1",
+            "task_config": {
+                "reuse_engine": true,
+                "aggregation": "max",
+                "models": [
+                    {
+                        "lang_code": "en",
+                        "model": "Babelscape/wikineural-multilingual-ner"
+                    },
+                    {
+                        "lang_code": "es",
+                        "model": "Babelscape/wikineural-multilingual-ner"
+                    },
+                    {
+                        "lang_code": "de",
+                        "model": "Babelscape/wikineural-multilingual-ner"
+                    },
+                    {
+                        "lang_code": "fr",
+                        "model": "Babelscape/wikineural-multilingual-ner"
+                    },
+                    {
+                        "lang_code": "it",
+                        "model": "Babelscape/wikineural-multilingual-ner"
+                    },
+                    {
+                        "lang_code": "pt",
+                        "model": "Babelscape/wikineural-multilingual-ner"
+                    }
+                ]
+            },
+            "pii_list": [
+                {
+                    "type": "PERSON",
+                    "lang": [
+                        "en",
+                        "es",
+                        "it",
+                        "pt",
+                        "de",
+                        "fr"
+                    ],
+                    "method": "model",
+                    "extra": {
+                        "map": "PER"
+                    }
+                },
+                {
+                    "type": "LOCATION",
+                    "lang": [
+                        "en",
+                        "es",
+                        "it",
+                        "pt",
+                        "de",
+                        "fr"
+                    ],
+                    "method": "model",
+                    "extra": {
+                        "map": "LOC"
+                    }
+                }
+            ]
+        },
+        {
+            "format": "piisa:config:pii-extract:plugins:v1",
+            "piisa-detectors-presidio": {
+                "load": false
+            }
+        }
+    ]
+}

examples.txt

@@ -0,0 +1,9 @@
+Anjali Mehra was born on July 15, 1987 in the bustling city of New York. She is an accomplished author and journalist who has written several best-selling novels and contributed to numerous publications. Anjali grew up in a small apartment in Manhattan with her parents, both of whom were immigrants from India. From a young age, she showed a keen interest in writing and storytelling, often regaling her family and friends with tales of adventure and intrigue. After graduating from Columbia University with a degree in English literature, Anjali began her career as a freelance writer for various magazines and websites. Her work quickly gained recognition, and she soon landed a job at one of the most prestigious publishing houses in the country. Over the years, Anjali has published several successful novels that have captivated readers around the world. Her latest book, "The Secret Garden," was released earlier this year and has already become a global sensation. In addition to her writing, Anjali is also known for her activism and philanthropy. She frequently speaks out against social justice issues and works tirelessly to raise awareness and funds for various causes. Despite her busy schedule, Anjali always makes time for her loved ones. She is married to fellow author, David Chen, and they have two children together. When she's not working or volunteering, Anjali can be found spending quality time with her family or exploring new places and cultures. If you want to get in touch with Anjali, you can reach her at her fake phone number (212) 555-1234 or her made-up address: 123 Main Street, New York, NY 10001.
+Rob and Jim Smith are a dynamic duo who have been making waves in the entertainment industry since their early days. Born on May 15, 1990, the brothers grew up in Los Angeles, California with their parents, both of whom were actors. From a young age, Rob and Jim showed a natural talent for acting and performing. They would often put on shows for their family and friends, much to everyone's delight. After high school, the brothers decided to pursue careers in Hollywood. Rob quickly rose to fame as a comedian, appearing in numerous stand-up routines and sketch comedy shows. He has also starred in several TV series and movies, including the popular sitcom "Modern Family" and the blockbuster film "The Hangover." Jim, on the other hand, took a different path and became a successful musician. He plays guitar and sings in his band, which has toured extensively throughout North America and Europe. Their music has been described as a fusion of rock, pop, and folk, and they have won critical acclaim for their unique sound. Despite their busy schedules, Rob and Jim make it a point to spend time together whenever possible. They enjoy going on adventures, trying new foods, and exploring new cities. If you want to get in touch with Rob and Jim, you can reach them at their fake phone numbers: Rob at (323) 555-5555 and Jim at (657) 555-5555, or their made-up address: 456 Elm Street, Los Angeles, CA 90031.
+Somsak Chou is a bright young man studying computer science at the National University of Laos. Born on January 1, 1995 in Vientiane, he has shown great promise in his academic pursuits and extracurricular activities. Somsak first discovered his passion for technology when he built his own website at the age of 16. Since then, he has continued to develop his skills and knowledge in web development, programming, and data analysis. Aside from his studies, Somsak is actively involved in various clubs and organizations at the university. He serves as president of the Computer Science Club and volunteers regularly at local schools and community centers. Somsak is fluent in English and enjoys learning about different cultures and languages. He hopes to use his education and skills to contribute positively to society and make a difference in the world. If you want to get in touch with Somsak, you can reach him at his fake email address: [email protected] or his fake phone number: +856 20 1234567. His made-up address is: 789 Moo 1, Tumbol Suthep, Amphoe Mueang Chiang Mai, Chang Wat Chiang Mai 50200, Thailand.
+Ana Rodriguez is a talented engineering student at the University of Seville in Spain. She was born on July 1, 1998 in Madrid and has always had a strong interest in math and science. Ana excelled in her studies and received a full scholarship to study engineering at the prestigious university in Seville. Ana is passionate about using her technical skills to solve real-world problems and improve people's lives. In addition to her coursework, she participates in various extracurricular activities such as volunteering at local hospitals and participating in robotics competitions. Ana speaks Spanish fluently and is eager to learn more about different cultures and languages. She hopes to one day work in an international company where she can apply her knowledge and skills to help others around the world. If you want to get in touch with Ana, you can reach her at her fake email address: [email protected] or her fake phone number: +34 987 654 321. Her made-up address is: Calle Real de la Alcazaba, 123, Sevilla, Andalucia, Spain.
+Jamie is a nonbinary individual who identifies outside of traditional gender categories. They were born on May 1, 1999 in New York City and have always felt comfortable expressing themselves in a way that does not conform to societal expectations. Jamie grew up surrounded by diverse communities and learned early on to embrace their unique identity. They have been active in LGBTQ+ activism since high school and continue to advocate for greater acceptance and understanding of nonbinary individuals. In addition to their advocacy efforts, Jamie enjoys exploring new hobbies and interests such as photography, writing, and cooking. They also value personal growth and self-care, often taking time off from social media and other digital distractions to focus on their mental health and wellbeing. Despite facing challenges related to their gender identity, Jamie remains optimistic about the future and hopes to see greater representation and inclusion for nonbinary individuals in all aspects of society. You can reach out to them via email at [email protected] or call them at (123) 456-7890.
+Esmeralda Gonzales es una persona ficticia que se describe como una mujer con un fuerte interés en el arte y la cultura. Nació en Madrid en 1985 y desde pequeña ha estado fascinada por los diferentes estilos artísticos y las tradiciones culturales del mundo entero. Esmeralda estudió pintura en la Universidad Complutense de Madrid y más tarde obtuvo su maestría en Historia del Arte en la Universidad Carlos III de Madrid. En sus años libres, ella viaja extensamente para explorar nuevos lugares y descubrir nuevas obras de arte. Además de ser una apasionada coleccionista de arte, Esmeralda también tiene un gran conocimiento sobre la historia y la cultura española. Ella ha publicado varios ensayos sobre temas relacionados con la historia y la cultura españolas y ha presentado numerosas conferencias en este campo. Si quieres ponerse en contacto con Esmeralda, puedes hacerlo mediante correo electrónico al [email protected] o llamarla al (+34) 987 654 321. Su dirección postal es: Calle Real de la Alcazaba, 123, Sevilla, Andalucía, España.
+Herr Schmidt ist ein deutscher Bauernmann, der seit seiner Kindheit auf dem Hof mit seinem Vater und Großvater die Landwirtschaft betreibt. Er wurde im Jahr 1960 geboren und hat sich eine tiefe Liebe für das Land und seine Tiere entwickelt. Herr Schmidt studierte Agronomie an der Universität Hohenheim und arbeitete als Fachberater für landwirtschaftliche Betriebe vor er sich selbst übernahm den Hof von seinen Eltern. Er produziert Obst, Getreide und Fleisch und verkauft sein Produkt an lokale Geschäfte und Restaurants. In seiner Freizeit genietet Herr Schmidt mit seiner Frau und ihren drei Kindern das Leben auf dem Hof. Sie lieben gemeinsame Aktivitäten wie Jagen, Angeln und Wanderungen durch die Natur. Wenn Sie Kontakt mit Herrn Schmidt aufnehmen möchten können Sie ihn per E-Mail kontaktieren unter [email protected] oder telefonieren Sie ihm an (+49) 123 456 7890. Sein Postanschrift ist: Schmidt Farm, Hauptstraße 123, Dorf XYZ, Bayern, Deutschland.
+Madame Marie est une boulangère française qui a découvert son passion pour la cuisine à un jeune âge. Elle a été née à Lyon en 1980 et a suivi des cours de formation professionnelle en boulangerie. Madame Marie a commencé sa carrière comme assistante dans une boulangerie avant d'ouvrir sa propre bakery dans le centre-ville de Lyon. Son lieu de vente propose une gamme variée de pains, croissants et desserts artisanalement réalisés à partir de farines et ingrédients locaux. Outre son activité professionnelle, Madame Marie aime partager ses connaissances culinaires avec ses amis et famille. Elle organise souvent des ateliers de cuisine et des repas communautaires. Si vous avez besoin d'en savoir plus sur ses produits ou si vous souhaitez passer commande, vous pouvez le contacter par courrier postal : [email protected] ou par téléphone : (+33) 4 56 78 90 12. Sa bakery est située à : Place Bouchard, n° 45, Lyon, France.
+Maria Gonçalves é uma marinheira portuguesa que tem sido apasionada pelo mar desde sua infância. Nasceu em Lisboa em 1990 e sempre teve um interesse particular por explorar os oceanos e descobrir novos lugares. Depois de terminar o ensino secundário, Maria decidiu seguir uma formação profissional na Escola Naval do Algarve para se tornar marinheira. Desde então, ela trabalhou em vários navios como oficial de comunicação e navegador. Em seus momentos livres, Maria gosta de praticar esportes aquáticos e participar em eventos náuticos. Também é uma grande admiradora da natureza e gosta de viajar pelos países europeus e as Américas. Se você quiser saber mais sobre Maria ou se pretender fazer uma visita ao seu navio, você pode contatar-lhe através do email: [email protected] ou telefonicamente: +351 912 345 678. Sua base naval está localizada no Porto de Lisboa, Portugal.

requirements.txt

@@ -0,0 +1,120 @@
+aiofiles @ file:///home/conda/feedstock_root/build_artifacts/aiofiles_1677173365185/work
+aiohttp @ file:///Users/runner/miniforge3/conda-bld/aiohttp_1676292777256/work
+aiosignal @ file:///home/conda/feedstock_root/build_artifacts/aiosignal_1667935791922/work
+altair @ file:///home/conda/feedstock_root/build_artifacts/altair_1683659205121/work
+anyio @ file:///home/conda/feedstock_root/build_artifacts/anyio_1666191106763/work/dist
+async-timeout @ file:///home/conda/feedstock_root/build_artifacts/async-timeout_1640026696943/work
+attrs @ file:///home/conda/feedstock_root/build_artifacts/attrs_1683424013410/work
+blis==0.7.9
+brotlipy @ file:///Users/runner/miniforge3/conda-bld/brotlipy_1666764759924/work
+catalogue==2.0.8
+cchardet==2.1.7
+certifi==2023.5.7
+cffi @ file:///Users/runner/miniforge3/conda-bld/cffi_1671179893800/work
+chardet==5.1.0
+charset-normalizer @ file:///tmp/build/80754af9/charset-normalizer_1630003229654/work
+click @ file:///home/conda/feedstock_root/build_artifacts/click_1666798198223/work
+colorama @ file:///home/conda/feedstock_root/build_artifacts/colorama_1666700638685/work
+confection==0.0.4
+contourpy @ file:///Users/runner/miniforge3/conda-bld/contourpy_1673633760692/work
+cryptography @ file:///Users/runner/miniforge3/conda-bld/cryptography-split_1681508747141/work
+cycler @ file:///home/conda/feedstock_root/build_artifacts/cycler_1635519461629/work
+cymem==2.0.7
+en-core-web-lg @ https://github.com/explosion/spacy-models/releases/download/en_core_web_lg-3.5.0/en_core_web_lg-3.5.0-py3-none-any.whl#sha256=c8ac64840c1eb3e3ca7bd38bd1e1c48fb0faeb2449d54d01d5ce629af4595775
+Faker==18.9.0
+fastapi @ file:///home/conda/feedstock_root/build_artifacts/fastapi_1684266487947/work
+ffmpy @ file:///home/conda/feedstock_root/build_artifacts/ffmpy_1659474992694/work
+filelock @ file:///home/conda/feedstock_root/build_artifacts/filelock_1681839547898/work
+fonttools @ file:///Users/runner/miniforge3/conda-bld/fonttools_1683740659111/work
+frozenlist @ file:///Users/runner/miniforge3/conda-bld/frozenlist_1667935502123/work
+fsspec @ file:///home/conda/feedstock_root/build_artifacts/fsspec_1683494881189/work
+gradio==3.46.1
+gradio_client==0.5.3
+h11 @ file:///home/conda/feedstock_root/build_artifacts/h11_1664132893548/work
+h2 @ file:///home/conda/feedstock_root/build_artifacts/h2_1634280454336/work
+hpack==4.0.0
+httpcore @ file:///home/conda/feedstock_root/build_artifacts/httpcore_1684362107262/work
+httpx @ file:///home/conda/feedstock_root/build_artifacts/httpx_1681225508856/work
+huggingface-hub==0.16.4
+hyperframe @ file:///home/conda/feedstock_root/build_artifacts/hyperframe_1619110129307/work
+idna @ file:///home/conda/feedstock_root/build_artifacts/idna_1663625384323/work
+importlib-metadata @ file:///home/conda/feedstock_root/build_artifacts/importlib-metadata_1682176699712/work
+importlib-resources @ file:///home/conda/feedstock_root/build_artifacts/importlib_resources_1676919000169/work
+Jinja2 @ file:///home/conda/feedstock_root/build_artifacts/jinja2_1654302431367/work
+jsonschema @ file:///home/conda/feedstock_root/build_artifacts/jsonschema-meta_1669810440410/work
+kiwisolver @ file:///Users/runner/miniforge3/conda-bld/kiwisolver_1666805765141/work
+langcodes==3.3.0
+linkify-it-py @ file:///home/conda/feedstock_root/build_artifacts/linkify-it-py_1651923627081/work
+lxml==4.9.2
+markdown-it-py @ file:///home/conda/feedstock_root/build_artifacts/markdown-it-py_1677100944732/work
+MarkupSafe @ file:///Users/runner/miniforge3/conda-bld/markupsafe_1674135859696/work
+matplotlib @ file:///Users/runner/miniforge3/conda-bld/matplotlib-suite_1678135763500/work
+mdit-py-plugins @ file:///home/conda/feedstock_root/build_artifacts/mdit-py-plugins_1670348296204/work
+mdurl @ file:///home/conda/feedstock_root/build_artifacts/mdurl_1639515908913/work
+mpmath==1.3.0
+multidict @ file:///Users/runner/miniforge3/conda-bld/multidict_1672339514000/work
+munkres==1.1.4
+murmurhash==1.0.9
+networkx==3.1
+numpy @ file:///Users/runner/miniforge3/conda-bld/numpy_1682210346059/work
+orjson @ file:///Users/runner/miniforge3/conda-bld/orjson_1683499478306/work/target/wheels/orjson-3.8.12-cp310-cp310-macosx_11_0_arm64.whl#sha256=6f78bf4d76dca252d8d6f48e062f4573cc99dd8b7ffff72f11ae03765f224733
+packaging @ file:///home/conda/feedstock_root/build_artifacts/packaging_1681337016113/work
+pandas @ file:///Users/runner/miniforge3/conda-bld/pandas_1683494188900/work
+pathy==0.10.1
+phonenumbers==8.13.11
+pii-data==0.4.0
+pii-extract-base==0.5.0
+pii-extract-plg-presidio==0.3.2
+pii-extract-plg-regex==0.4.1
+pii-extract-plg-transformers==0.1.1
+pii-preprocess==0.0.4
+pii-transform==0.5.1
+Pillow @ file:///Users/runner/miniforge3/conda-bld/pillow_1680694547271/work
+pkgutil_resolve_name @ file:///home/conda/feedstock_root/build_artifacts/pkgutil-resolve-name_1633981968097/work
+preshed==3.0.8
+presidio-analyzer==2.2.32
+pycparser @ file:///home/conda/feedstock_root/build_artifacts/pycparser_1636257122734/work
+pydantic @ file:///Users/runner/miniforge3/conda-bld/pydantic_1679565524285/work
+pydub @ file:///home/conda/feedstock_root/build_artifacts/pydub_1615612442567/work
+pyOpenSSL @ file:///home/conda/feedstock_root/build_artifacts/pyopenssl_1680037383858/work
+pyparsing @ file:///home/conda/feedstock_root/build_artifacts/pyparsing_1652235407899/work
+pyrsistent @ file:///Users/runner/miniforge3/conda-bld/pyrsistent_1672681537313/work
+PySocks @ file:///home/conda/feedstock_root/build_artifacts/pysocks_1661604839144/work
+python-dateutil @ file:///home/conda/feedstock_root/build_artifacts/python-dateutil_1626286286081/work
+python-docx==0.8.11
+python-multipart @ file:///home/conda/feedstock_root/build_artifacts/python-multipart_1679167423335/work
+python-stdnum==1.18
+pytz @ file:///home/conda/feedstock_root/build_artifacts/pytz_1680088766131/work
+PyYAML @ file:///Users/runner/miniforge3/conda-bld/pyyaml_1666772543719/work
+regex==2023.5.5
+requests @ file:///home/conda/feedstock_root/build_artifacts/requests_1682535435083/work
+requests-file==1.5.1
+safetensors==0.3.3
+semantic-version @ file:///home/conda/feedstock_root/build_artifacts/semantic_version_1653579368137/work
+six @ file:///home/conda/feedstock_root/build_artifacts/six_1620240208055/work
+smart-open==6.3.0
+sniffio @ file:///home/conda/feedstock_root/build_artifacts/sniffio_1662051266223/work
+spacy==3.5.3
+spacy-legacy==3.0.12
+spacy-loggers==1.0.4
+srsly==2.4.6
+starlette @ file:///home/conda/feedstock_root/build_artifacts/starlette-recipe_1684245096404/work
+sympy==1.12
+thinc==8.1.10
+tldextract==3.4.3
+tokenizers==0.14.0
+toolz @ file:///home/conda/feedstock_root/build_artifacts/toolz_1657485559105/work
+torch==2.1.0
+tqdm @ file:///home/conda/feedstock_root/build_artifacts/tqdm_1677948868469/work
+transformers==4.34.0
+typer==0.7.0
+typing_extensions @ file:///home/conda/feedstock_root/build_artifacts/typing_extensions_1678559861143/work
+tzdata @ file:///home/conda/feedstock_root/build_artifacts/python-tzdata_1680081134351/work
+uc-micro-py @ file:///home/conda/feedstock_root/build_artifacts/uc-micro-py_1608058642472/work
+unicodedata2 @ file:///Users/runner/miniforge3/conda-bld/unicodedata2_1667239979860/work
+urllib3 @ file:///home/conda/feedstock_root/build_artifacts/urllib3_1678635778344/work
+uvicorn @ file:///Users/runner/miniforge3/conda-bld/uvicorn-split_1682667915083/work
+wasabi==1.1.1
+websockets @ file:///Users/runner/miniforge3/conda-bld/websockets_1683550497847/work
+yarl @ file:///Users/runner/miniforge3/conda-bld/yarl_1682426738484/work
+zipp @ file:///home/conda/feedstock_root/build_artifacts/zipp_1677313463193/work