Merge branch 'release/1.2.0'

.circleci/config.yml

@@ -0,0 +1,106 @@
+.common-values:
+
+  docker-image: &docker-image circleci/python:3.6.4
+
+  restore-cache: &restore-cache
+    keys:
+      - v1-dependencies-{{ checksum "setup.py" }}-{{ checksum "requirements.txt" }}
+
+  create-venv: &create-venv
+    name: Create virtualenv
+    command: /usr/local/bin/python3 -m venv venv
+
+  save-cache: &save-cache
+    paths:
+      - ./venv
+    key: v1-dependencies-{{ checksum "setup.py" }}-{{ checksum "requirements.txt" }}
+
+  install-package: &install-package
+    name: Install package
+    command: |
+      . venv/bin/activate
+      venv/bin/python3 -m pip install tensorflow==1.13.1
+      venv/bin/python3 -m pip install .
+
+version: 2
+jobs:
+  build:
+    docker:
+      - image: *docker-image
+    steps:
+    - checkout
+    - restore_cache: *restore-cache
+    - run: *create-venv
+    - save_cache: *save-cache
+    - run: *install-package
+
+  pip-check:
+    docker:
+      - image: *docker-image
+    steps:
+    - checkout
+    - restore_cache: *restore-cache
+    - run: *create-venv
+    - save_cache: *save-cache
+    - run: *install-package
+    - run:
+        name: Pip check
+        command: |
+          . venv/bin/activate
+          venv/bin/python3 -m pip check
+
+  test:
+    docker:
+      - image: *docker-image
+    steps:
+    - checkout
+    - restore_cache: *restore-cache
+    - run: *create-venv
+    - save_cache: *save-cache
+    - run: *install-package
+    - run:
+          name: Unit tests with Pytest
+          command: |
+            . venv/bin/activate
+            venv/bin/python3 -m pytest tests
+  lint:
+    docker:
+      - image: *docker-image
+    steps:
+    - checkout
+    - restore_cache: *restore-cache
+    - run: *create-venv
+    - save_cache: *save-cache
+    - run:
+          name: Linting with flake8
+          command: |
+            . venv/bin/activate
+            venv/bin/python3 -m pip install flake8
+            venv/bin/python3 -m flake8 ampligraph --max-line-length 120 --ignore=W291,W293
+
+  docs:
+    docker:
+      - image: *docker-image
+    steps:
+    - checkout
+    - restore_cache: *restore-cache
+    - run: *create-venv
+    - save_cache: *save-cache
+    - run: *install-package
+    - run:
+          name: Making docs with Sphinx
+          command: |
+            . venv/bin/activate
+            cd docs
+            make clean autogen html
+
+
+workflows:
+  version: 2
+  checks:
+    jobs:
+      - build
+      - pip-check
+      - lint
+      - docs
+      - test

README.md

@@ -4,6 +4,8 @@
 
 [![Documentation Status](https://readthedocs.org/projects/ampligraph/badge/?version=latest)](http://ampligraph.readthedocs.io/?badge=latest)
 
+[Join the conversation on Slack](https://join.slack.com/t/ampligraph/shared_invite/enQtNTc2NTI0MzUxMTM5LTRkODk0MjI2OWRlZjdjYmExY2Q3M2M3NGY0MGYyMmI4NWYyMWVhYTRjZDhkZjA1YTEyMzBkMGE4N2RmNTRiZDg)
+![](docs/img/slack_logo.png)
 
 **Open source library based on TensorFlow that predicts links between concepts in a knowledge graph.**
 
@@ -41,7 +43,7 @@ It then combines embeddings with model-specific scoring functions to predict uns
 AmpliGraph includes the following submodules:
 
 * **Datasets**: helper functions to load datasets (knowledge graphs).
-* **Models**: knowledge graph embedding models. AmpliGraph contains TransE, DistMult, ComplEx, HolE. (More to come!)
+* **Models**: knowledge graph embedding models. AmpliGraph contains **TransE**, **DistMult**, **ComplEx**, **HolE**, **ConvKB**. (More to come!)
 * **Evaluation**: metrics and evaluation protocols to assess the predictive power of the models.
 * **Discovery**: High-level convenience APIs for knowledge discovery (discover new facts, cluster entities, predict near duplicates).
 
@@ -114,24 +116,24 @@ pip install -e .
 ```python
 >> import ampligraph
 >> ampligraph.__version__
-'1.1.0'
+'1.2.0'
 ```
 
 
 ## Predictive Power Evaluation (MRR Filtered)
 
-AmpliGraph includes implementations of TransE, DistMult, ComplEx and HolE. 
+AmpliGraph includes implementations of TransE, DistMult, ComplEx, HolE and ConvKB.
 Their predictive power is reported below and compared against the state-of-the-art results in literature.
 [More details available here](https://docs.ampligraph.org/en/latest/experiments.html).
 
-|                 |FB15K-237 |WN18RR   |YAGO3-10   | FB15k      |WN18           |
-|-----------------|----------|---------|-----------|------------|---------------|
-| Literature Best | **0.35***| 0.48*   | 0.49*     | **0.84**** | **0.95***     |
-| TransE (AmpliGraph)         |  0.31    | 0.22    | 0.49      | 0.63       | 0.65          |
-| DistMult (AmpliGraph)        |  0.31    | 0.45    | 0.49      | 0.78       | 0.82          |
-| ComplEx  (AmpliGraph)       |  0.32    | **0.50**| **0.50**  | 0.80       | 0.94          |
-| HolE (AmpliGraph)           |  0.31    | 0.47    | **0.50**  | 0.80       | 0.93          |
-
+|                              |FB15K-237 |WN18RR   |YAGO3-10   | FB15k      |WN18           |
+|------------------------------|----------|---------|-----------|------------|---------------|
+| Literature Best              | **0.35***| 0.48*   | 0.49*     | **0.84**** | **0.95***     |
+| TransE (AmpliGraph)          |  0.31    | 0.22    | **0.51**  | 0.63       | 0.66          |
+| DistMult (AmpliGraph)        |  0.31    | 0.47    | 0.50      | 0.78       | 0.82          |
+| ComplEx  (AmpliGraph)        |  0.32    | **0.51**| 0.49      | 0.80       | 0.94          |
+| HolE (AmpliGraph)            |  0.31    | 0.47    | 0.50      | 0.80       | 0.94          |
+| ConvKB (AmpliGraph)          |  0.23    | 0.39    | 0.30      | 0.65       | 0.80          |
 
 <sub>
 * Timothee Lacroix, Nicolas Usunier, and Guillaume Obozinski. Canonical tensor decomposition for knowledge base 

ampligraph/__init__.py

@@ -9,7 +9,7 @@
 import logging.config
 import pkg_resources
 
-__version__ = '1.1.0'
+__version__ = '1.2.0'
 __all__ = ['datasets', 'latent_features', 'discovery', 'evaluation', 'utils']
 
 logging.config.fileConfig(pkg_resources.resource_filename(__name__, 'logger.conf'), disable_existing_loggers=False)

ampligraph/datasets/__init__.py

@@ -8,13 +8,13 @@
 """Helper functions to load knowledge graphs."""
 
 from .datasets import load_from_csv, load_from_rdf, load_fb15k, load_wn18, load_fb15k_237, load_from_ntriples, \
-    load_yago3_10, load_wn18rr
+    load_yago3_10, load_wn18rr, load_wn11, load_fb13
 
 from .abstract_dataset_adapter import AmpligraphDatasetAdapter
 from .sqlite_adapter import SQLiteAdapter
 from .numpy_adapter import NumpyDatasetAdapter
 
 
 __all__ = ['load_from_csv', 'load_from_rdf', 'load_from_ntriples', 'load_wn18', 'load_fb15k',
-           'load_fb15k_237', 'load_yago3_10', 'load_wn18rr', 
+           'load_fb15k_237', 'load_yago3_10', 'load_wn18rr', 'load_wn11', 'load_fb13',
            'AmpligraphDatasetAdapter', 'NumpyDatasetAdapter', 'SQLiteAdapter']

ampligraph/datasets/abstract_dataset_adapter.py

@@ -89,55 +89,26 @@ def set_filter(self, filter_triples):
         """
         raise NotImplementedError('Abstract Method not implemented!')
 
-    def get_next_train_batch(self, batch_size=1, dataset_type="train"):
+    def get_next_batch(self, batches_count=-1, dataset_type="train", use_filter=False):
         """Generator that returns the next batch of data.
         
         Parameters
         ----------
-        batch_size : int
-            data size that needs to be returned
-        dataset_type: string
-            indicates which dataset to use
-        Returns
-        -------
-        batch_output : nd-array
-            yields a batch of triples from the dataset type specified
-        """
-        raise NotImplementedError('Abstract Method not implemented!')
-
-    def get_next_eval_batch(self, batch_size=1, dataset_type="test"):
-        """Generator that returns the next batch of data.
-        
-        Parameters
-        ----------
-        batch_size : int
-            data size that needs to be returned
         dataset_type: string
             indicates which dataset to use
+        batches_count: int
+            number of batches per epoch (default: -1, i.e. uses batch_size of 1)
+        use_filter : bool
+            Flag to indicate whether to return the concepts that need to be filtered
+
         Returns
         -------
         batch_output : nd-array
             yields a batch of triples from the dataset type specified
-        """
-        raise NotImplementedError('Abstract Method not implemented!')
-
-    def get_next_batch_with_filter(self, batch_size=1, dataset_type="test"):
-        """Generator that returns the next batch of data along with the filter.
-        
-        Parameters
-        ----------
-        batch_size : int
-            data size that needs to be returned
-        dataset_type: string
-            indicates which dataset to use
-        Returns
-        -------
-        batch_output : nd-array [n,3]
-            yields a batch of triples from the dataset type specified
         participating_objects : nd-array [n,1]
-            all objects that were involved in the s-p-? relation
+            all objects that were involved in the s-p-? relation. This is returned only if use_filter is set to true.
         participating_subjects : nd-array [n,1]
-            all subjects that were involved in the ?-p-o relation
+            all subjects that were involved in the ?-p-o relation. This is returned only if use_filter is set to true.
         """
         raise NotImplementedError('Abstract Method not implemented!')