initial commit

.gitignore

@@ -0,0 +1 @@
+.env

Dockerfile

@@ -0,0 +1,27 @@
+FROM python:3.9
+
+#COPY requirements.txt /app/
+WORKDIR /app
+
+RUN --mount=type=secret,id=LICENSE_KEY,mode=0444,required=true \
+    pip install --upgrade pip \
+    && pip install prodigy -f https://$(cat /run/secrets/LICENSE_KEY)@download.prodi.gy
+#    && pip install --trusted-host pypi.python.org -r requirements.txt
+
+# COPY wheel/prodigy-xxx-linux_x86_64.whl ./wheel/
+# RUN pip install wheel/prodigy-xxx-linux_x86_64.whl \
+#     && rm -rf wheel/prodigy-xxx-linux_x86_64.whl
+RUN python -m spacy download en_core_web_sm
+
+COPY prodigy.json .
+COPY data ./data/
+
+ENV PRODIGY_HOME /app
+ENV PRODIGY_LOGGING "verbose"
+ENV PRODIGY_ALLOWED_SESSIONS "user1,user2"
+ENV PRODIGY_BASIC_AUTH_USER "admin"
+ENV PRODIGY_BASIC_AUTH_PASS "password"
+
+EXPOSE 7860
+
+CMD python -m prodigy ner.manual ner_news en_core_web_sm ./data/dataset.jsonl --label PERSON,ORG,PRODUCT

data/dataset.jsonl

@@ -0,0 +1,200 @@
+{"text":"Uber\u2019s Lesson: Silicon Valley\u2019s Start-Up Machine Needs Fixing","meta":{"source":"The New York Times"}}
+{"text":"Pearl Automation, Founded by Apple Veterans, Shuts Down","meta":{"source":"The New York Times"}}
+{"text":"How Silicon Valley Pushed Coding Into American Classrooms","meta":{"source":"The New York Times"}}
+{"text":"Women in Tech Speak Frankly on Culture of Harassment","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley Investors Flexed Their Muscles in Uber Fight","meta":{"source":"The New York Times"}}
+{"text":"Uber is a Creature of an Industry Struggling to Grow Up","meta":{"source":"The New York Times"}}
+{"text":"\u2018The Internet Is Broken\u2019: @ev Is Trying to Salvage It","meta":{"source":"The New York Times"}}
+{"text":"The South Park Commons Fills a Hole in the Tech Landscape","meta":{"source":"The New York Times"}}
+{"text":"The Closing of the Republican Mind","meta":{"source":"The New York Times"}}
+{"text":"Writers From the Right and Left on Trump Jr., the Future of the F.B.I., Health Care and More","meta":{"source":"The New York Times"}}
+{"text":"Daily Report: From Lean to Fat Start-Ups","meta":{"source":"The New York Times"}}
+{"text":"How Uber\u2019s Chief Is Gaining Even More Clout in the Company","meta":{"source":"The New York Times"}}
+{"text":"As New Zealand Courts Tech Talent, Isolation Becomes a Draw","meta":{"source":"The New York Times"}}
+{"text":"One Thing Silicon Valley Can\u2019t Seem to Fix","meta":{"source":"The New York Times"}}
+{"text":"In Silicon Valley, a Voice of Caution Guides a High-Flying Uber","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley Writes a Protest Letter Against Trump","meta":{"source":"The New York Times"}}
+{"text":"Warriors, Tech\u2019s Team, Are Soaring Out of Reach","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley\u2019s Most Elusive Beast","meta":{"source":"The New York Times"}}
+{"text":"Wall Street and Silicon Valley Form an Uneasy Alliance","meta":{"source":"The New York Times"}}
+{"text":"Tim O'Reilly Explains the Internet of Things","meta":{"source":"The New York Times"}}
+{"text":"Alibaba I.P.O. May Unleash Global Fight Over Users","meta":{"source":"The New York Times"}}
+{"text":"Disruptions: Looking Beyond Silicon Valley's Bubble","meta":{"source":"The New York Times"}}
+{"text":"Twitter Outages Linked to Glitches and Site Upgrade","meta":{"source":"The New York Times"}}
+{"text":"Out of the Loop in Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"A Determined Outpost of Tiny Technology","meta":{"source":"The New York Times"}}
+{"text":"In Silicon Valley, Recruiters Are Sending Out Their Own R\u00e9sum\u00e9s","meta":{"source":"The New York Times"}}
+{"text":"Credit Crisis Spreads a Pall Over Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"As Department Stores Close, Stitch Fix Expands Online","meta":{"source":"The New York Times"}}
+{"text":"Tech Incubators on a Mission of Diversity","meta":{"source":"The New York Times"}}
+{"text":"Meal-Delivery Start-Ups Take Aim at Your Dinner","meta":{"source":"The New York Times"}}
+{"text":"If a Bubble Bursts in Palo Alto, Does It Make a Sound?","meta":{"source":"The New York Times"}}
+{"text":"Web\u2019s Reach Binds N.S.A. and Silicon Valley Leaders","meta":{"source":"The New York Times"}}
+{"text":"Daily Report: Other Nations Offer Visas to Compete With Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"Bill Maher Concert Will Introduce New Yahoo! Comedy Channel","meta":{"source":"The New York Times"}}
+{"text":"A Silicon Valley Dream Grows in Guatemala, Despite the Risks","meta":{"source":"The New York Times"}}
+{"text":"One Country, Two Revolutions","meta":{"source":"The New York Times"}}
+{"text":"The Class That Built Apps, and Fortunes","meta":{"source":"The New York Times"}}
+{"text":"E-Ties That Bind","meta":{"source":"The New York Times"}}
+{"text":"Tech Recruiting Clashes With Immigration Rules","meta":{"source":"The New York Times"}}
+{"text":"On Day Care, Google Makes a Rare Fumble","meta":{"source":"The New York Times"}}
+{"text":"Where Computer Artifacts Come Alive","meta":{"source":"The New York Times"}}
+{"text":"Can Green Make Green?","meta":{"source":"The New York Times"}}
+{"text":"In Silicon Valley, the Crash Seems Like Just Yesterday","meta":{"source":"The New York Times"}}
+{"text":"\u2018I\u2019m Here to Help,\u2019 Trump Tells Tech Executives at Meeting","meta":{"source":"The New York Times"}}
+{"text":"Peter Thiel\u2019s Bet on Donald Trump Pays Off","meta":{"source":"The New York Times"}}
+{"text":"Peter Thiel Defends His Most Contrarian Move Yet: Supporting Trump","meta":{"source":"The New York Times"}}
+{"text":"What It\u2019s Like to Fight Online Hate","meta":{"source":"The New York Times"}}
+{"text":"Start-Ups Once Showered With Cash Now Have to Work for It","meta":{"source":"The New York Times"}}
+{"text":"European Tech Scene Begins to Feel Silicon Valley\u2019s Woes","meta":{"source":"The New York Times"}}
+{"text":"Toyota Invests $1 Billion in Artificial Intelligence in U.S.","meta":{"source":"The New York Times"}}
+{"text":"Farhad and Mike\u2019s Week in Review: Twitter Problems","meta":{"source":"The New York Times"}}
+{"text":"Steep Discounts a Boon for Customers, but a Gamble for Start-Ups","meta":{"source":"The New York Times"}}
+{"text":"Apple's Diversity Mirrors Other Tech Companies'   ","meta":{"source":"The New York Times"}}
+{"text":"Building a Better Battery","meta":{"source":"The New York Times"}}
+{"text":"Disruptions: If It Looks Like a Bubble and Floats Like a Bubble ...","meta":{"source":"The New York Times"}}
+{"text":"Disruptions: Even the Tech Elites Leave Gadgets Behind","meta":{"source":"The New York Times"}}
+{"text":"Latest Product From Tech Firms: An Immigration Bill","meta":{"source":"The New York Times"}}
+{"text":"Migrant Bill Seems to Fit Tech Sector Wish List","meta":{"source":"The New York Times"}}
+{"text":"The Internet Gets Physical","meta":{"source":"The New York Times"}}
+{"text":"The War Between Apple and Google Has Just Begun","meta":{"source":"The New York Times"}}
+{"text":"Site Lets Investors See and Copy Experts\u2019 Trades","meta":{"source":"The New York Times"}}
+{"text":"Amid Conference Halls and Keynote Speakers, a Rivalry Forms","meta":{"source":"The New York Times"}}
+{"text":"Economy Has Become a Drag on Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"Yahoo Decides to Release a Rosy Forecast","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley Losing Middle-Wage Jobs","meta":{"source":"The New York Times"}}
+{"text":"The Working-Class Millionaire","meta":{"source":"The New York Times"}}
+{"text":"Andy Grove\u2019s Warning to Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"Caution Rebuffed, Unicorns and Other Start-Ups Fixate on Rainbows","meta":{"source":"The New York Times"}}
+{"text":"Kleiner Perkins, Disrupted","meta":{"source":"The New York Times"}}
+{"text":"When Uber and Airbnb Meet the Real World","meta":{"source":"The New York Times"}}
+{"text":"Yahoo Reveals Work Force Data, Joining Tech's Small Diversity Parade","meta":{"source":"The New York Times"}}
+{"text":"Disruptions: The Echo Chamber of Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"TimesCast Tech: Career and Family in Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"Steve Jobs: He Brought the Show to Business","meta":{"source":"The New York Times"}}
+{"text":"For Start-Ups, Late Bursts of Private Cash","meta":{"source":"The New York Times"}}
+{"text":"AOL, Seeking to Revitalize Its Internet Role, Turns to a Candid Ex-Yahoo Executive","meta":{"source":"The New York Times"}}
+{"text":"Buy.com Deal With eBay Angers Sellers","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley Starts to Turn Its Face to the Sun","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley\u2019s High-Tech Hunt for Colleague","meta":{"source":"The New York Times"}}
+{"text":"Twitter\u2019s Troubles and Snap\u2019s Appeal: It\u2019s All About the Mojo","meta":{"source":"The New York Times"}}
+{"text":"A Rarity at a Republican Convention: \u2018I Am Proud to Be Gay\u2019","meta":{"source":"The New York Times"}}
+{"text":"Pinterest Hires Its First Head of Diversity","meta":{"source":"The New York Times"}}
+{"text":"Daily Report: Narendra Modi, Indian Prime Minister, Conquers Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"Inside Amazon: Reporter\u2019s Notebook","meta":{"source":"The New York Times"}}
+{"text":"Hot Tech Start-Ups May Face a Long and Bumpy Fall","meta":{"source":"The New York Times"}}
+{"text":"White House Takes Cybersecurity Pitch to Silicon Valley ","meta":{"source":"The New York Times"}}
+{"text":"The PayPal Mafia\u2019s Golden Touch","meta":{"source":"The New York Times"}}
+{"text":"What We're Reading","meta":{"source":"The New York Times"}}
+{"text":"Will Boom Lead to Bust in Silicon Valley? ","meta":{"source":"The New York Times"}}
+{"text":"Tech Industry Flexes Muscle in California Race","meta":{"source":"The New York Times"}}
+{"text":"Disruptions: A Blogger Mocks the Denizens of Silicon Valley ","meta":{"source":"The New York Times"}}
+{"text":"How Big Data Is Playing Recruiter for Specialized Workers","meta":{"source":"The New York Times"}}
+{"text":"Daily Report: Multinationals Push Into Venture Capital","meta":{"source":"The New York Times"}}
+{"text":"As Silicon Valley Cheers Yahoo Chief, Wall Street's Reaction Is Muted","meta":{"source":"The New York Times"}}
+{"text":"Old Techies Never Die; They Just Can\u2019t Get Hired as an Industry Moves On","meta":{"source":"The New York Times"}}
+{"text":"A Silicon Bubble Shows Signs of Reinflating","meta":{"source":"The New York Times"}}
+{"text":"A Start-Up Says It Can Predict Others\u2019 Fate","meta":{"source":"The New York Times"}}
+{"text":"More Than Games, a Net to Snare Social Networkers","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley Shaped by Technology and Traffic","meta":{"source":"The New York Times"}}
+{"text":"In Silicon Valley, Millionaires Who Don\u2019t Feel Rich","meta":{"source":"The New York Times"}}
+{"text":"Lots of Froth but No Bubble","meta":{"source":"The New York Times"}}
+{"text":"Start-Up Fervor Shifts to Energy in Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"VeriSign Moves to Address an Internet Security Problem","meta":{"source":"The New York Times"}}
+{"text":"China, Not Silicon Valley, Is Cutting Edge in Mobile Tech","meta":{"source":"The New York Times"}}
+{"text":"Women in Tech Band Together to Track Diversity, After Hours","meta":{"source":"The New York Times"}}
+{"text":"Facebook\u2019s Developer Conference Kicks Off","meta":{"source":"The New York Times"}}
+{"text":"An Anonymous Satire of Silicon Valley Now Has a Publisher","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley Bank Strengthens Its Roots","meta":{"source":"The New York Times"}}
+{"text":"Daily Report: Rivalry Brewing Between Uber and Google","meta":{"source":"The New York Times"}}
+{"text":"Behind the Cover Story: Yiren Lu on the Angst, Perks and Failures of the Tech Industry","meta":{"source":"The New York Times"}}
+{"text":"Thorny Side Effects in Silicon Valley Tactic to Keep Control","meta":{"source":"The New York Times"}}
+{"text":"Is New York's Tech Boom Sustainable?","meta":{"source":"The New York Times"}}
+{"text":"Internet Giants Foster, and Threaten, Innovation Economy","meta":{"source":"The New York Times"}}
+{"text":"Following Venture Capital for Signs of Tech to Come","meta":{"source":"The New York Times"}}
+{"text":"At Social Site, Only the Businesslike Need Apply","meta":{"source":"The New York Times"}}
+{"text":"When It Comes to Innovation, Geography Is Destiny","meta":{"source":"The New York Times"}}
+{"text":"Tech Billionaire Backing Trump Suggests Silicon Valley Is Out of Touch","meta":{"source":"The New York Times"}}
+{"text":"How Silicon Valley Treats a Trump Backer: Peter Thiel","meta":{"source":"The New York Times"}}
+{"text":"Parents Ready for Some Love From Silicon Valley Companies","meta":{"source":"The New York Times"}}
+{"text":"Scrutiny of Security Start-Ups May Signal Shift in Venture Funding","meta":{"source":"The New York Times"}}
+{"text":"Marissa Mayer of Yahoo Says She’s Pregnant With Twins","meta":{"source":"The New York Times"}}
+{"text":"A Novel Prompts a Conversation About How We Use Technology","meta":{"source":"The New York Times"}}
+{"text":"Tech Industry Sets Its Sights on Gambling","meta":{"source":"The New York Times"}}
+{"text":"Techies Break a Fashion Taboo","meta":{"source":"The New York Times"}}
+{"text":"Back to the Future: The Netscape and Google I.P.O.'s","meta":{"source":"The New York Times"}}
+{"text":"Where Are the Women Executives in Silicon Valley?","meta":{"source":"The New York Times"}}
+{"text":"For Buyers of Web Start-Ups, Quest to Corral Young Talent","meta":{"source":"The New York Times"}}
+{"text":"Digital Muse for Beat Poet","meta":{"source":"The New York Times"}}
+{"text":"Amid Conference Halls and Keynote Speakers, a Rivalry Forms","meta":{"source":"The New York Times"}}
+{"text":"Hostility Has Its Rewards","meta":{"source":"The New York Times"}}
+{"text":"Going Public Caps Dream for a Maker of Software","meta":{"source":"The New York Times"}}
+{"text":"Next Job for Obama? Silicon Valley Is Hiring","meta":{"source":"The New York Times"}}
+{"text":"Airbnb and Others Set Terms for Employees to Cash Out","meta":{"source":"The New York Times"}}
+{"text":"Google and Apple: the High-Tech Hippies of Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley, Seeking Diversity, Focuses on Blacks","meta":{"source":"The New York Times"}}
+{"text":"As More Tech Start-Ups Stay Private, So Does the Money","meta":{"source":"The New York Times"}}
+{"text":"Want a Steady Income? There\u2019s an App for That","meta":{"source":"The New York Times"}}
+{"text":"Investing Early On for Insights, Not Profits","meta":{"source":"The New York Times"}}
+{"text":"Angel Investors Lend Expertise as Well as Cash","meta":{"source":"The New York Times"}}
+{"text":"Deal Makers Invade CES, the Land of Geeks","meta":{"source":"The New York Times"}}
+{"text":"Techies Break a Fashion Taboo","meta":{"source":"The New York Times"}}
+{"text":"Back to the Future: The Netscape and Google I.P.O.'s","meta":{"source":"The New York Times"}}
+{"text":"Where Are the Women Executives in Silicon Valley?","meta":{"source":"The New York Times"}}
+{"text":"For Buyers of Web Start-Ups, Quest to Corral Young Talent","meta":{"source":"The New York Times"}}
+{"text":"Digital Muse for Beat Poet","meta":{"source":"The New York Times"}}
+{"text":"Amid Conference Halls and Keynote Speakers, a Rivalry Forms","meta":{"source":"The New York Times"}}
+{"text":"Hostility Has Its Rewards","meta":{"source":"The New York Times"}}
+{"text":"Going Public Caps Dream for a Maker of Software","meta":{"source":"The New York Times"}}
+{"text":"Next Job for Obama? Silicon Valley Is Hiring","meta":{"source":"The New York Times"}}
+{"text":"Airbnb and Others Set Terms for Employees to Cash Out","meta":{"source":"The New York Times"}}
+{"text":"In Surveillance Debate, White House Turns Its Focus to Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley\u2019s Youth Problem","meta":{"source":"The New York Times"}}
+{"text":"Zuckerberg Remains the Undisputed Boss at Facebook","meta":{"source":"The New York Times"}}
+{"text":"A Corporate Campus Made to Mirror Facebook","meta":{"source":"The New York Times"}}
+{"text":"One on One: Tim Wu, Author of 'The Master Switch'","meta":{"source":"The New York Times"}}
+{"text":"Spinning the Web: P.R. in Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"To Survive, Net Start-Ups Slow Their Metabolism","meta":{"source":"The New York Times"}}
+{"text":"Attending to the Needs of the Too-Busy","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley Start-Ups Awash in Dollars, Again","meta":{"source":"The New York Times"}}
+{"text":"New Social Sites Cater to People of a Certain Age","meta":{"source":"The New York Times"}}
+{"text":"In Surveillance Debate, White House Turns Its Focus to Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley\u2019s Youth Problem","meta":{"source":"The New York Times"}}
+{"text":"Zuckerberg Remains the Undisputed Boss at Facebook","meta":{"source":"The New York Times"}}
+{"text":"A Corporate Campus Made to Mirror Facebook","meta":{"source":"The New York Times"}}
+{"text":"One on One: Tim Wu, Author of 'The Master Switch'","meta":{"source":"The New York Times"}}
+{"text":"Spinning the Web: P.R. in Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"To Survive, Net Start-Ups Slow Their Metabolism","meta":{"source":"The New York Times"}}
+{"text":"Attending to the Needs of the Too-Busy","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley Start-Ups Awash in Dollars, Again","meta":{"source":"The New York Times"}}
+{"text":"New Social Sites Cater to People of a Certain Age","meta":{"source":"The New York Times"}}
+{"text":"Demand for Data Puts Engineers in Spotlight","meta":{"source":"The New York Times"}}
+{"text":"Yahoo Sale Could Be Bad for Minnows","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley Helped Create Trump, and That\u2019s Bad for It","meta":{"source":"The New York Times"}}
+{"text":"Peter Thiel\u2019s Embrace of Trump Has Silicon Valley Squirming","meta":{"source":"The New York Times"}}
+{"text":"In Silicon Valley, a Divide in Income and Politics","meta":{"source":"The New York Times"}}
+{"text":"The Art of Failing Upward","meta":{"source":"The New York Times"}}
+{"text":"Daily Report: The Tech News Cycle is About to Ramp Up","meta":{"source":"The New York Times"}}
+{"text":"Netromancy","meta":{"source":"The New York Times"}}
+{"text":"Unicorns Hunt for Talent Among Silicon Valley\u2019s Giants","meta":{"source":"The New York Times"}}
+{"text":"Overvalued in Silicon Valley, but Don\u2019t Say \u2018Tech Bubble\u2019","meta":{"source":"The New York Times"}}
+{"text":"Managers Turn to Computer Games, Aiming for More Efficient Employees","meta":{"source":"The New York Times"}}
+{"text":"Stanford Women and Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"What It Means to Be a \u2018Dad\u2019","meta":{"source":"The New York Times"}}
+{"text":"What We're Reading","meta":{"source":"The New York Times"}}
+{"text":"A Golden Age of Design","meta":{"source":"The New York Times"}}
+{"text":"Court Rejects Deal on Hiring in Silicon Valley","meta":{"source":"The New York Times"}}
+{"text":"What the Beats Deal Says About Apple: It Loves Tastemakers","meta":{"source":"The New York Times"}}
+{"text":"Search for a Market Niche, and You Might Find a Crowd","meta":{"source":"The New York Times"}}
+{"text":"Search for the 'Next Big Thing' Yields Soaring Valuations","meta":{"source":"The New York Times"}}
+{"text":"Shifting Tech Scene Unsettles Big Players","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley\u2019s Start-Up Machine","meta":{"source":"The New York Times"}}
+{"text":"Opening a Gateway for Girls to Enter the Computer Field","meta":{"source":"The New York Times"}}
+{"text":"Silicon Valley Says Step Away From the Device","meta":{"source":"The New York Times"}}
+{"text":"Like the Apple Store Nearby, but This One Has Windows","meta":{"source":"The New York Times"}}
+{"text":"A King of Connections Is Tech\u2019s Go-To Guy","meta":{"source":"The New York Times"}}
+{"text":"One on One: Jaron Lanier","meta":{"source":"The New York Times"}}
+{"text":"6 Months, $90,000 and (Maybe) a Great Idea","meta":{"source":"The New York Times"}}
+{"text":"Microsoft\u2019s Failed Yahoo Bid Risks Online Growth","meta":{"source":"The New York Times"}}
+{"text":"Why Old Technologies Are Still Kicking","meta":{"source":"The New York Times"}}
+{"text":"Trying to Add a Pulse to a World of Machines","meta":{"source":"The New York Times"}}

prodigy.json

@@ -0,0 +1,6 @@
+{
+    "port": 7860,
+    "host": "0.0.0.0",
+    "feed_overlap": true,
+    "show_stats": true
+}