Add linting rules for tfjs-models. (#333)

- Adds lint rules to mirror monorepo
- Fixes lint errors
- Adds CI to run lint
- Refactors body pix and posenet to simplify base models for resnet / mobilenet.

Author: Nikhil Thorat

.tslint/noImportsFromDistRule.js

@@ -0,0 +1,58 @@
+'use strict';
+var __extends = (this && this.__extends) || (function() {
+                  var extendStatics = function(d, b) {
+                    extendStatics = Object.setPrototypeOf ||
+                        ({__proto__: []} instanceof Array && function(d, b) {
+                                      d.__proto__ = b;
+                                    }) || function(d, b) {
+                          for (var p in b)
+                            if (b.hasOwnProperty(p)) d[p] = b[p];
+                        };
+                    return extendStatics(d, b);
+                  };
+                  return function(d, b) {
+                    extendStatics(d, b);
+                    function __() {
+                      this.constructor = d;
+                    }
+                    d.prototype = b === null ?
+                        Object.create(b) :
+                        (__.prototype = b.prototype, new __());
+                  };
+                })();
+exports.__esModule = true;
+var Lint = require('tslint');
+var Rule = /** @class */ (function(_super) {
+  __extends(Rule, _super);
+  function Rule() {
+    return _super !== null && _super.apply(this, arguments) || this;
+  }
+  Rule.prototype.apply = function(sourceFile) {
+    return this.applyWithWalker(
+        new NoImportsFromDistWalker(sourceFile, this.getOptions()));
+  };
+  Rule.FAILURE_STRING =
+      'importing from dist/ is prohibited. Please use public API';
+  return Rule;
+}(Lint.Rules.AbstractRule));
+exports.Rule = Rule;
+var NoImportsFromDistWalker = /** @class */ (function(_super) {
+  __extends(NoImportsFromDistWalker, _super);
+  function NoImportsFromDistWalker() {
+    return _super !== null && _super.apply(this, arguments) || this;
+  }
+  NoImportsFromDistWalker.prototype.visitImportDeclaration = function(node) {
+    var importFrom = node.moduleSpecifier.getText();
+    var reg = /@tensorflow\/tfjs[-a-z]*\/dist/;
+    if (importFrom.match(reg)) {
+      var fix = new Lint.Replacement(
+          node.moduleSpecifier.getStart(), node.moduleSpecifier.getWidth(),
+          importFrom.replace(/\/dist[\/]*/, ''));
+      this.addFailure(this.createFailure(
+          node.moduleSpecifier.getStart(), node.moduleSpecifier.getWidth(),
+          Rule.FAILURE_STRING, fix));
+    }
+    _super.prototype.visitImportDeclaration.call(this, node);
+  };
+  return NoImportsFromDistWalker;
+}(Lint.RuleWalker));

.tslint/noImportsFromDistRule.ts

@@ -0,0 +1,30 @@
+import * as Lint from 'tslint';
+import * as ts from 'typescript';
+
+export class Rule extends Lint.Rules.AbstractRule {
+  public static FAILURE_STRING =
+      'importing from dist/ is prohibited. Please use public API';
+
+  public apply(sourceFile: ts.SourceFile): Lint.RuleFailure[] {
+    return this.applyWithWalker(
+        new NoImportsFromDistWalker(sourceFile, this.getOptions()));
+  }
+}
+
+class NoImportsFromDistWalker extends Lint.RuleWalker {
+  public visitImportDeclaration(node: ts.ImportDeclaration) {
+    const importFrom = node.moduleSpecifier.getText();
+    const reg = /@tensorflow\/tfjs[-a-z]*\/dist/;
+    if (importFrom.match(reg)) {
+      const fix = new Lint.Replacement(
+          node.moduleSpecifier.getStart(), node.moduleSpecifier.getWidth(),
+          importFrom.replace(/\/dist[\/]*/, ''));
+
+      this.addFailure(this.createFailure(
+          node.moduleSpecifier.getStart(), node.moduleSpecifier.getWidth(),
+          Rule.FAILURE_STRING, fix));
+    }
+
+    super.visitImportDeclaration(node);
+  }
+}

body-pix/cloudbuild.yml

@@ -14,6 +14,14 @@ steps:
   args: ['install']
   waitFor: ['yarn-common']
 
+# Lint.
+- name: 'node:10'
+  dir: 'body-pix'
+  entrypoint: 'yarn'
+  id: 'lint'
+  args: ['lint']
+  waitFor: ['yarn']
+
 # Build.
 - name: 'node:10'
   dir: 'body-pix'

body-pix/demos/package.json

@@ -18,7 +18,8 @@
   "scripts": {
     "watch": "cross-env NODE_ENV=development parcel index.html --no-hmr --open ",
     "build": "cross-env NODE_ENV=production parcel build index.html --public-url ./",
-    "lint": "eslint ."
+    "lint": "eslint .",
+    "link-local": "yalc link"
   },
   "browser": {
     "crypto": false

body-pix/demos/yarn.lock

@@ -6243,11 +6243,6 @@ typedarray@^0.0.6:
   resolved "https://registry.yarnpkg.com/typedarray/-/typedarray-0.0.6.tgz#867ac74e3864187b1d3d47d996a78ec5c8830777"
   integrity sha1-hnrHTjhkGHsdPUfZlqeOxciDB3c=
 
-[email protected]:
-  version "0.0.54"
-  resolved "https://registry.yarnpkg.com/typeface-oswald/-/typeface-oswald-0.0.54.tgz#1e253011622cdd50f580c04e7d625e7f449763d7"
-  integrity sha512-U1WMNp4qfy4/3khIfHMVAIKnNu941MXUfs3+H9R8PFgnoz42Hh9pboSFztWr86zut0eXC8byalmVhfkiKON/8Q==
-
 uncss@^0.17.0:
   version "0.17.2"
   resolved "https://registry.yarnpkg.com/uncss/-/uncss-0.17.2.tgz#fac1c2429be72108e8a47437c647d58cf9ea66f1"

body-pix/package.json

@@ -34,7 +34,7 @@
     "rollup-plugin-typescript2": "~0.13.0",
     "rollup-plugin-uglify": "~3.0.0",
     "ts-node": "~5.0.0",
-    "tslint": "~5.8.0",
+    "tslint": "~5.18.0",
     "typescript": "~3.5.3",
     "yalc": "^1.0.0-pre.27"
   },

body-pix/src/base_model.ts

@@ -0,0 +1,113 @@
+
+/**
+ * @license
+ * Copyright 2019 Google Inc. All Rights Reserved.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ * =============================================================================
+ */
+
+import * as tfconv from '@tensorflow/tfjs-converter';
+import * as tf from '@tensorflow/tfjs-core';
+import {BodyPixOutputStride} from './types';
+
+/**
+ * BodyPix supports using various convolution neural network models
+ * (e.g. ResNet and MobileNetV1) as its underlying base model.
+ * The following BaseModel interface defines a unified interface for
+ * creating such BodyPix base models. Currently both MobileNet (in
+ * ./mobilenet.ts) and ResNet (in ./resnet.ts) implements the BaseModel
+ * interface. New base models that conform to the BaseModel interface can be
+ * added to BodyPix.
+ */
+export abstract class BaseModel {
+  constructor(
+      protected readonly model: tfconv.GraphModel,
+      public readonly outputStride: BodyPixOutputStride) {
+    const inputShape =
+        this.model.inputs[0].shape as [number, number, number, number];
+    tf.util.assert(
+        (inputShape[1] === -1) && (inputShape[2] === -1),
+        () => `Input shape [${inputShape[1]}, ${inputShape[2]}] ` +
+            `must both be equal to or -1`);
+  }
+
+  abstract preprocessInput(input: tf.Tensor3D): tf.Tensor3D;
+
+  /**
+   * Predicts intermediate Tensor representations.
+   *
+   * @param input The input RGB image of the base model.
+   * A Tensor of shape: [`inputResolution`, `inputResolution`, 3].
+   *
+   * @return A dictionary of base model's intermediate predictions.
+   * The returned dictionary should contains the following elements:
+   * - heatmapScores: A Tensor3D that represents the keypoint heatmap scores.
+   * - offsets: A Tensor3D that represents the offsets.
+   * - displacementFwd: A Tensor3D that represents the forward displacement.
+   * - displacementBwd: A Tensor3D that represents the backward displacement.
+   * - segmentation: A Tensor3D that represents the segmentation of all
+   * people.
+   * - longOffsets: A Tensor3D that represents the long offsets used for
+   * instance grouping.
+   * - partHeatmaps: A Tensor3D that represents the body part segmentation.
+   */
+  predict(input: tf.Tensor3D): {
+    heatmapScores: tf.Tensor3D,
+    offsets: tf.Tensor3D,
+    displacementFwd: tf.Tensor3D,
+    displacementBwd: tf.Tensor3D,
+    segmentation: tf.Tensor3D,
+    partHeatmaps: tf.Tensor3D,
+    longOffsets: tf.Tensor3D,
+    partOffsets: tf.Tensor3D
+  } {
+    return tf.tidy(() => {
+      const asFloat = this.preprocessInput(input.toFloat());
+      const asBatch = asFloat.expandDims(0);
+      const results = this.model.predict(asBatch) as tf.Tensor4D[];
+      const results3d: tf.Tensor3D[] = results.map(y => y.squeeze([0]));
+      const namedResults = this.nameOutputResults(results3d);
+
+      return {
+        heatmapScores: namedResults.heatmap.sigmoid(),
+        offsets: namedResults.offsets,
+        displacementFwd: namedResults.displacementFwd,
+        displacementBwd: namedResults.displacementBwd,
+        segmentation: namedResults.segmentation,
+        partHeatmaps: namedResults.partHeatmaps,
+        longOffsets: namedResults.longOffsets,
+        partOffsets: namedResults.partOffsets
+      };
+    });
+  }
+
+  // Because MobileNet and ResNet predict() methods output a different order for
+  // these values, we have a method that needs to be implemented to order them.
+  abstract nameOutputResults(results: tf.Tensor3D[]): {
+    heatmap: tf.Tensor3D,
+    offsets: tf.Tensor3D,
+    displacementFwd: tf.Tensor3D,
+    displacementBwd: tf.Tensor3D,
+    segmentation: tf.Tensor3D,
+    partHeatmaps: tf.Tensor3D,
+    longOffsets: tf.Tensor3D,
+    partOffsets: tf.Tensor3D
+  };
+
+  /**
+   * Releases the CPU and GPU memory allocated by the model.
+   */
+  dispose() {
+    this.model.dispose();
+  }
+}

body-pix/src/body_pix_model.ts

@@ -19,56 +19,19 @@
 import * as tfconv from '@tensorflow/tfjs-converter';
 import * as tf from '@tensorflow/tfjs-core';
 
+import {BaseModel} from './base_model';
 import {decodeOnlyPartSegmentation, decodePartSegmentation, toMaskTensor} from './decode_part_map';
-import {MobileNet, MobileNetMultiplier} from './mobilenet';
+import {MobileNet} from './mobilenet';
 import {decodePersonInstanceMasks, decodePersonInstancePartMasks} from './multi_person/decode_instance_masks';
 import {decodeMultiplePoses} from './multi_person/decode_multiple_poses';
 import {ResNet} from './resnet';
 import {mobileNetSavedModel, resNet50SavedModel} from './saved_models';
-import {decodeSinglePose} from './sinlge_person/decode_single_pose';
+import {decodeSinglePose} from './single_person/decode_single_pose';
 import {BodyPixArchitecture, BodyPixInput, BodyPixInternalResolution, BodyPixMultiplier, BodyPixOutputStride, BodyPixQuantBytes, Padding, PartSegmentation, PersonSegmentation} from './types';
 import {getInputSize, padAndResizeTo, scaleAndCropToInputTensorShape, scaleAndFlipPoses, toTensorBuffers3D, toValidInternalResolutionNumber} from './util';
 
-
 const APPLY_SIGMOID_ACTIVATION = true;
 
-/**
- * BodyPix supports using various convolution neural network models
- * (e.g. ResNet and MobileNetV1) as its underlying base model.
- * The following BaseModel interface defines a unified interface for
- * creating such BodyPix base models. Currently both MobileNet (in
- * ./mobilenet.ts) and ResNet (in ./resnet.ts) implements the BaseModel
- * interface. New base models that conform to the BaseModel interface can be
- * added to BodyPix.
- */
-export interface BaseModel {
-  // The output stride of the base model.
-  readonly outputStride: BodyPixOutputStride;
-
-  /**
-   * Predicts intermediate Tensor representations.
-   *
-   * @param input The input RGB image of the base model.
-   * A Tensor of shape: [`inputResolution`, `inputResolution`, 3].
-   *
-   * @return A dictionary of base model's intermediate predictions.
-   * The returned dictionary should contains the following elements:
-   * - heatmapScores: A Tensor3D that represents the keypoint heatmap scores.
-   * - offsets: A Tensor3D that represents the offsets.
-   * - displacementFwd: A Tensor3D that represents the forward displacement.
-   * - displacementBwd: A Tensor3D that represents the backward displacement.
-   * - segmentation: A Tensor3D that represents the segmentation of all people.
-   * - longOffsets: A Tensor3D that represents the long offsets used for
-   * instance grouping.
-   * - partHeatmaps: A Tensor3D that represents the body part segmentation.
-   */
-  predict(input: tf.Tensor3D): {[key: string]: tf.Tensor3D};
-  /**
-   * Releases the CPU and GPU memory allocated by the model.
-   */
-  dispose(): void;
-}
-
 /**
  * BodyPix model loading is configurable using the following config dictionary.
  *
@@ -101,7 +64,7 @@ export interface BaseModel {
 export interface ModelConfig {
   architecture: BodyPixArchitecture;
   outputStride: BodyPixOutputStride;
-  multiplier?: MobileNetMultiplier;
+  multiplier?: BodyPixMultiplier;
   modelUrl?: string;
   quantBytes?: BodyPixQuantBytes;
 }
@@ -602,15 +565,15 @@ export class BodyPix {
       };
     });
 
-    const [scoresBuffer, offsetsBuffer, displacementsFwdBuffer, displacementsBwdBuffer] =
+    const [scoresBuf, offsetsBuf, displacementsFwdBuf, displacementsBwdBuf] =
         await toTensorBuffers3D([
           heatmapScoresRaw, offsetsRaw, displacementFwdRaw, displacementBwdRaw
         ]);
 
-    let poses = await decodeMultiplePoses(
-        scoresBuffer, offsetsBuffer, displacementsFwdBuffer,
-        displacementsBwdBuffer, this.baseModel.outputStride,
-        config.maxDetections, config.scoreThreshold, config.nmsRadius);
+    let poses = decodeMultiplePoses(
+        scoresBuf, offsetsBuf, displacementsFwdBuf, displacementsBwdBuf,
+        this.baseModel.outputStride, config.maxDetections,
+        config.scoreThreshold, config.nmsRadius);
 
     poses = scaleAndFlipPoses(
         poses, [height, width],
@@ -849,15 +812,15 @@ export class BodyPix {
       };
     });
 
-    const [scoresBuffer, offsetsBuffer, displacementsFwdBuffer, displacementsBwdBuffer] =
+    const [scoresBuf, offsetsBuf, displacementsFwdBuf, displacementsBwdBuf] =
         await toTensorBuffers3D([
           heatmapScoresRaw, offsetsRaw, displacementFwdRaw, displacementBwdRaw
         ]);
 
-    let poses = await decodeMultiplePoses(
-        scoresBuffer, offsetsBuffer, displacementsFwdBuffer,
-        displacementsBwdBuffer, this.baseModel.outputStride,
-        config.maxDetections, config.scoreThreshold, config.nmsRadius);
+    let poses = decodeMultiplePoses(
+        scoresBuf, offsetsBuf, displacementsFwdBuf, displacementsBwdBuf,
+        this.baseModel.outputStride, config.maxDetections,
+        config.scoreThreshold, config.nmsRadius);
 
     poses = scaleAndFlipPoses(
         poses, [height, width],