add pages to docs

effectai · May 1, 2024 · 8df5d79 · 8df5d79
1 parent 980f045
commit 8df5d79
Show file tree

Hide file tree

Showing 18 changed files with 602 additions and 192 deletions.
diff --git a/docs/pages/docs/collecting-data/adding-tasks.mdx b/docs/pages/docs/collecting-data/adding-tasks.mdx
@@ -0,0 +1,30 @@
+## Adding tasks
+
+Adding tasks to a campaign is done through <u>batches</u> batches are a collection of tasks that are added to a campaign. Each batch can contain multiple tasks.
+Let's start by creating a batch with 3 tasks to our newly created image classification campaign.
+
+```ts [example.ts]
+import { createBatch } from '@effectai/effect-js'
+
+const { actor } = client.session
+
+const batch = await createBatch({
+  client,
+  batch: {
+      campaign_id: '<..>', // the id of the campaign
+      repetitions: 1 // number of times each task should be repeated
+      payer: actor // the actor that will pay the workers for the tasks
+  }
+  data: [
+    {
+      image: 'https://example.com/image.jpg', //task 1 image placeholder
+    },
+    {
+      image: 'https://example.com/image2.jpg', //task 2 image placeholder
+    },
+    {
+      image: 'https://example.com/image3.jpg', // task 3 image placeholder
+    }
+  ]
+})
+```
diff --git a/docs/pages/docs/collecting-data/create-a-campaign.mdx b/docs/pages/docs/collecting-data/create-a-campaign.mdx
@@ -0,0 +1,43 @@
+## Creating Your First Campaign
+
+Campaigns are the main way to collect data on Effect Network.
+A campaign is a <u>collection of tasks</u> that need to be completed by workers. Each task is a small piece of work that needs to be done. For example, a task could be to label an image, transcribe a piece of audio, or answer a question.
+ They are created by the data requester (you) and contain information like the reward for the workers, the template for the tasks, the instructions for the workers etc.
+
+### Creating a campaign
+
+```ts [example.ts]
+import { createCampaign } from '@effectai/effect-js'
+
+// get the actor from the session
+const { actor } = client.session
+
+// example template
+const template = '<html>...</html>'
+
+const campaign = await createCampaign({
+  client,
+  data: { 
+      title: 'My first campaign',
+      category: 'Image Labeling', // Category of the campaign e.g. Image Labeling, Audio Transcription etc.
+      description: 'This is a description of my first campaign', // Description of the campaign
+      estimated_time: 3600, // Estimated time to complete a task in seconds
+      example_task: 'This is an example task',
+      image: 'https://example.com/image.png',
+      instructions: 'This is the instructions for the workers',
+      input_schema: null,
+      output_schema: null,
+      template,
+      version: 1,
+  },
+  campaign: {
+    owner: actor,
+    reward: { quantity: '1', contract: 'effecttokens' },
+    max_task_time: 3600,
+    qualis: [],
+    payer: actor,
+  },
+})
+```
+
+Wooooho! You have created your first campaign. You can now view the campaign on the [Effect Network](https://app.effect.network/campaigns) before you start collecting data, you need to add tasks to the campaign. You can do this by following the [Adding Tasks](/docs/collecting-data/add-tasks) guide.
diff --git a/docs/pages/docs/collecting-data/create-a-template.mdx b/docs/pages/docs/collecting-data/create-a-template.mdx
@@ -0,0 +1,57 @@
+
+# Create a Template
+Before we can create a data collection campaign, we need to create a **template**.
+
+A template is a blueprint for the tasks in a campaign. It defines the structure of the tasks, the input and output data.
+Templates are created by the data requester (you) and can be reused across multiple campaigns.
+
+A typical template will be a HTML file that contains the structure of the tasks. The template can contain placeholders for the input data, output data, and other dynamic content.
+
+For this guide, we will create a simple template for an image labeling campaign. The template contains a title, an image, and a question for the workers to answer. The image is loaded from IPFS using the `ipfs` placeholder.
+
+```html
+<html>
+  <head>
+    <title>My Image labeling Campaign</title>
+  </head>
+  <body>
+    <form>
+        <div class="classify-image">
+            <h1 class="title">Classify The Image</h1>
+
+              <!-- The image loaded from IPFS  -->
+            <img src='https://gateway.pinata.cloud/ipfs/${ipfs}'/>
+        </div>
+        <div class="question question-sentiment">
+              <!-- Our question to ask the workers-->
+            <h2 class="subtitle">How much does this resemble the original one?</h2>
+
+              <!-- Input options for the workers-->
+            <div class="field">
+                <input class="is-checkradio" id="full" value="full" type="radio" name="resemblance" required>
+                <label for="full">Completly</label>
+                <input class="is-checkradio" id="half" value="half" type="radio" name="resemblance">
+                <label for="half">Half</label>
+                <input class="is-checkradio" id="none" value="none" type="radio" name="resemblance">
+                <label for="none">Nothing At All</label>
+            </div>
+
+        </div>
+
+        <!-- Submit button so workers can submit their data -->
+        <button type="submit" class="is-large mt-6">Submit</button>
+      </form>
+  </body>
+
+<script>
+setTimeout(() => console.log(window.FORCE_PLACEHOLDERS), 1000)
+</script>
+
+</html>
+```
+
+You can find examples for different types of templates on our [github](https://github.com/effectai/effect-force-templates)
+We highly recommended to follow the [template guide](/docs/templates/introduction) when you are creating your own template.
+
+
+Now that we have created a template for our campaign, we can move on to the next step: [Creating a Campaign](/docs/collecting-data/create-a-campaign)
diff --git a/docs/pages/docs/collecting-data/introduction.mdx b/docs/pages/docs/collecting-data/introduction.mdx
diff --git a/docs/pages/docs/gettingStarted.mdx → docs/pages/docs/getting-started.mdx b/docs/pages/docs/gettingStarted.mdx → docs/pages/docs/getting-started.mdx
@@ -1,7 +1,5 @@
 # Getting Started [Get started with the SDK in just a few lines of code.]
 
-## Overview
-
 ## Installation
 
 Use your favorite package manager to install the SDK.

diff --git a/docs/pages/docs/guides/createACampaign.mdx b/docs/pages/docs/guides/createACampaign.mdx
diff --git a/docs/pages/docs/introduction.mdx b/docs/pages/docs/introduction.mdx
@@ -1,57 +1,14 @@
 # Why Effect AI
 
-## Introduction
-
-Effect AI was built to solve the problem of AI development.
-It's a platform that connects AI developers with AI trainers to create data to train AI models.
-
 ## Problems
 
-The problem that Effect AI is trying to solve is the lack of training data for AI models.
-AI models are only as good as the data they are trained on.
-The more data you have, the better the model will be.
-
-There is an issue though that now we are facing, AI models are being trained with data that ws created by other AI models.
-This is a problem because the data is not always accurate, the model will learn and exacerbate mistakes from the training data.
-Leading to unreliable AI models.
+AI will be the pivotal technology of the future that will transcend us into a new era of efficiency and innovation.
+However, the development of AI models is contingent upon the availability of high-quality training data.
+This poses a significant challenge, as the creation of such data is a labor-intensive and time-consuming process.
+Moreover, this data is often gathered & developed behind closed doors, leading to a lack of transparency and accountability.
 
 ## Solution
 
-The soluiton is to create high quality training data.
-This is where Effect AI comes in.
-
-Effect AI is a platform that connects AI developers with AI trainers.
-
-AI developers can post tasks that they need training data for.
-Ai trainers can then complete these tasks and get paid for their work.
-
----
-
-
-In the rapidly evolving landscape of artificial intelligence (AI), the quest for optimal development and performance remains paramount.
-Effect AI emerges as a beacon in this realm, offering a solution to a pressing dilemma: the scarcity of high-quality training data for AI models.
-
-At its core, Effect AI addresses the fundamental challenge of AI development—the insufficiency of robust datasets.
-As the adage goes, "garbage in, garbage out"; AI models are only as proficient as the data they are trained upon.
-Consequently, the quest for superior data becomes imperative for fostering the accuracy and reliability of AI technologies.
-
-However, a disconcerting trend has emerged in the AI domain—models being trained with data generated by other AI models.
-This recursive process, while ostensibly efficient, precipitates a cascade of issues.
-Foremost among these concerns is the potential for error amplification: inaccuracies within the training data can be perpetuated and magnified within the AI model, resulting in flawed outcomes and unreliable performance.
-
-Enter Effect AI, a platform meticulously crafted to address these challenges.
-Serving as a nexus between AI developers and AI trainers, Effect AI facilitates the creation of high-quality training data, thereby fortifying the foundations of AI development.
-
-The modus operandi of Effect AI is elegantly simple yet profoundly impactful.
-AI developers leverage the platform to articulate their data requirements through the formulation of tasks.
-These tasks, ranging from image classification to natural language processing, encapsulate the diverse array of challenges encountered in AI development.
-
-In response, AI trainers—individuals proficient in data annotation and curation—step forward to fulfill these tasks.
-Through their collective efforts, they meticulously craft datasets that adhere to the highest standards of quality and accuracy.
-In doing so, they contribute not merely to the advancement of AI technologies but to the cultivation of a virtuous cycle wherein superior data begets superior models.
-
-Crucially, Effect AI's platform incentivizes participation and excellence.
-AI trainers are duly compensated for their contributions, thereby fostering a symbiotic relationship wherein both developers and trainers are mutually benefited.
-
-In conclusion, Effect AI stands as a testament to the power of innovation in addressing complex challenges.
-By championing the creation of high-quality training data, it not only mitigates the pitfalls of current AI development practices but also charts a course towards a future where AI technologies are synonymous with reliability, accuracy, and societal benefit.
+The solution is Effect AI, a decentralized network that connects AI developers with a global workforce of AI trainers.
+Leading the charge in the democratization of AI, Effect AI enables developers to create high-quality training data through a transparent and collaborative process.
+By leveraging the power of the crowd, Effect AI ensures that AI models are trained on diverse and reliable datasets, thereby enhancing their accuracy and performance.
diff --git a/docs/pages/docs/localDevelopment.mdx → docs/pages/docs/local-development.mdx b/docs/pages/docs/localDevelopment.mdx → docs/pages/docs/local-development.mdx
diff --git a/docs/pages/docs/templates/introduction.mdx b/docs/pages/docs/templates/introduction.mdx
@@ -0,0 +1,6 @@
+# Introduction
+
+
+## What is a template?
+
+## Creating Your own template
diff --git a/docs/pages/docs/templates/placeholders.mdx b/docs/pages/docs/templates/placeholders.mdx
diff --git a/docs/pages/index.mdx b/docs/pages/index.mdx
@@ -10,7 +10,14 @@ logo: false
 import { HomePage } from 'vocs/components'
 
 
-<div className="absolute max-w-[1120px] mx-auto w-full top-20 bottom-0 bg-[url('/colosseum-light.svg')] dark:bg-[url('/colosseum.svg')] bg-no-repeat max-sm:w-[200%] max-sm:-left-[200px] max-sm:hidden" />
+<div
+style={{
+  "background": 
+  `linear-gradient(180deg, rgba(255, 255, 255, 0.5) 0%, rgba(255, 255, 255, 1) 100%),
+  url('/preview-bg.png')
+  `
+}}
+ className="absolute mx-auto w-full top-20 bottom-0 bg-no-repeat max-sm:w-[200%] max-sm:-left-[200px] max-sm:hidden bg-cover left-0" />
 
 <div className="max-w-[1120px] mx-auto vp-doc relative px-[24px] mb-[96px] mt-[32px] md:px-0 md:mb-[64px]">
   <div className="pt-[48px] max-sm:pt-0">
@@ -19,7 +26,7 @@ import { HomePage } from 'vocs/components'
         <img className="h-[72px] dark:invert-[1] max-sm:h-[60px]" src="/logo.png" alt="Effect logo" />
         <div className="font-regular text-[21px] max-sm:text-[18px] text-[#919193] max-md:text-center">Effect AI is the data network for training next-gen transparent AI models. Integrate the SDK into your app and tap into a global, decentralized workforce.</div>
         <div className="flex justify-center space-x-2">
-          <HomePage.Button href="/docs/gettingStarted" variant="accent">Get started</HomePage.Button>
+          <HomePage.Button href="/docs/getting-started" variant="accent">Get started</HomePage.Button>
           <HomePage.Button href="https://github.com/effectai">GitHub</HomePage.Button>
         </div>
       </div>

diff --git a/docs/public/preview-bg.png b/docs/public/preview-bg.png