Update tutorials (torch versions, ES version, replace Finder with Pipeline) (#814)

* remove manual torch install on colab

* update elasticsearch version everywhere to 7.9.2

* fix FAQPipeline

* update tutorials with new pipelines

* Add latest docstring and tutorial changes

* revert faqpipeline change. fix field names in tutorial 4

* Add latest docstring and tutorial changes

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

docs/_src/tutorials/tutorials/1.md

@@ -44,13 +44,11 @@ Make sure you enable the GPU runtime to experience decent speed in this tutorial
 # Install the latest master of Haystack
 !pip install git+https://github.com/deepset-ai/haystack.git
 !pip install urllib3==1.25.4
-!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html
 
 ```
 
 
 ```python
-from haystack import Finder
 from haystack.preprocessor.cleaning import clean_wiki_text
 from haystack.preprocessor.utils import convert_files_to_dicts, fetch_archive_from_http
 from haystack.reader.farm import FARMReader
@@ -74,19 +72,19 @@ You can start Elasticsearch on your local machine instance using Docker. If Dock
 
 ```python
 # Recommended: Start Elasticsearch using Docker
-#! docker run -d -p 9200:9200 -e "discovery.type=single-node" elasticsearch:7.6.2
+#! docker run -d -p 9200:9200 -e "discovery.type=single-node" elasticsearch:7.9.2
 ```
 
 
 ```python
 # In Colab / No Docker environments: Start Elasticsearch from source
-! wget https://artifacts.elastic.co/downloads/elasticsearch/elasticsearch-7.6.2-linux-x86_64.tar.gz -q
+! wget https://artifacts.elastic.co/downloads/elasticsearch/elasticsearch-7.9.2-linux-x86_64.tar.gz -q
-! tar -xzf elasticsearch-7.6.2-linux-x86_64.tar.gz
+! tar -xzf elasticsearch-7.9.2-linux-x86_64.tar.gz
-! chown -R daemon:daemon elasticsearch-7.6.2
+! chown -R daemon:daemon elasticsearch-7.9.2
 
 import os
 from subprocess import Popen, PIPE, STDOUT
-es_server = Popen(['elasticsearch-7.6.2/bin/elasticsearch'],
+es_server = Popen(['elasticsearch-7.9.2/bin/elasticsearch'],
                    stdout=PIPE, stderr=STDOUT,
                    preexec_fn=lambda: os.setuid(1)  # as daemon
                   )
@@ -206,13 +204,17 @@ reader = FARMReader(model_name_or_path="deepset/roberta-base-squad2", use_gpu=Tr
 # reader = TransformersReader(model_name_or_path="distilbert-base-uncased-distilled-squad", tokenizer="distilbert-base-uncased", use_gpu=-1)
 ```
 
-### Finder
+### Pipeline
 
-The Finder sticks together reader and retriever in a pipeline to answer our actual questions. 
+With a Haystack `Pipeline` you can stick together your building blocks to a search pipeline.
+Under the hood, `Pipelines` are Directed Acyclic Graphs (DAGs) that you can easily customize for your own use cases.
+To speed things up, Haystack also comes with a few predefined Pipelines. One of them is the `ExtractiveQAPipeline` that combines a retriever and a reader to answer our questions.
+You can learn more about `Pipelines` in the [docs](https://haystack.deepset.ai/docs/latest/pipelinesmd).
 
 
 ```python
-finder = Finder(reader, retriever)
+from haystack.pipeline import ExtractiveQAPipeline
+pipe = ExtractiveQAPipeline(reader, retriever)
 ```
 
 ## Voilà! Ask a question!
@@ -221,13 +223,13 @@ finder = Finder(reader, retriever)
 ```python
 # You can configure how many candidates the reader and retriever shall return
 # The higher top_k_retriever, the better (but also the slower) your answers. 
-prediction = finder.get_answers(question="Who is the father of Arya Stark?", top_k_retriever=10, top_k_reader=5)
+prediction = pipe.run(query="Who is the father of Arya Stark?", top_k_retriever=10, top_k_reader=5)
 ```
 
 
 ```python
-# prediction = finder.get_answers(question="Who created the Dothraki vocabulary?", top_k_reader=5)
+# prediction = pipe.run(query="Who created the Dothraki vocabulary?", top_k_reader=5)
-# prediction = finder.get_answers(question="Who is the sister of Sansa?", top_k_reader=5)
+# prediction = pipe.run(query="Who is the sister of Sansa?", top_k_reader=5)
 ```
 
 

docs/_src/tutorials/tutorials/2.md

@@ -39,8 +39,6 @@ Make sure you enable the GPU runtime to experience decent speed in this tutorial
 # Install the latest master of Haystack
 !pip install git+https://github.com/deepset-ai/haystack.git
 !pip install urllib3==1.25.4
-!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html
-
 ```
 
 

docs/_src/tutorials/tutorials/3.md

@@ -39,8 +39,6 @@ Make sure you enable the GPU runtime to experience decent speed in this tutorial
 # Install the latest master of Haystack
 !pip install git+https://github.com/deepset-ai/haystack.git
 !pip install urllib3==1.25.4
-!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html
-
 ```
 
 
@@ -153,13 +151,17 @@ reader = FARMReader(model_name_or_path="deepset/roberta-base-squad2", use_gpu=Tr
 # reader = TransformersReader(model_name_or_path="distilbert-base-uncased-distilled-squad", tokenizer="distilbert-base-uncased", use_gpu=-1)
 ```
 
-### Finder
+### Pipeline
 
-The Finder sticks together reader and retriever in a pipeline to answer our actual questions. 
+With a Haystack `Pipeline` you can stick together your building blocks to a search pipeline.
+Under the hood, `Pipelines` are Directed Acyclic Graphs (DAGs) that you can easily customize for your own use cases.
+To speed things up, Haystack also comes with a few predefined Pipelines. One of them is the `ExtractiveQAPipeline` that combines a retriever and a reader to answer our questions.
+You can learn more about `Pipelines` in the [docs](https://haystack.deepset.ai/docs/latest/pipelinesmd).
 
 
 ```python
-finder = Finder(reader, retriever)
+from haystack.pipeline import ExtractiveQAPipeline
+pipe = ExtractiveQAPipeline(reader, retriever)
 ```
 
 ## Voilà! Ask a question!
@@ -167,14 +169,14 @@ finder = Finder(reader, retriever)
 
 ```python
 # You can configure how many candidates the reader and retriever shall return
-# The higher top_k_retriever, the better (but also the slower) your answers. 
+# The higher top_k_retriever, the better (but also the slower) your answers.
-prediction = finder.get_answers(question="Who is the father of Arya Stark?", top_k_retriever=10, top_k_reader=5)
+prediction = pipe.run(query="Who is the father of Arya Stark?", top_k_retriever=10, top_k_reader=5)
 ```
 
 
 ```python
-# prediction = finder.get_answers(question="Who created the Dothraki vocabulary?", top_k_reader=5)
+# prediction = pipe.run(query="Who created the Dothraki vocabulary?", top_k_reader=5)
-# prediction = finder.get_answers(question="Who is the sister of Sansa?", top_k_reader=5)
+# prediction = pipe.run(query="Who is the sister of Sansa?", top_k_reader=5)
 ```
 
 

docs/_src/tutorials/tutorials/4.md

@@ -47,8 +47,6 @@ Make sure you enable the GPU runtime to experience decent speed in this tutorial
 # Install the latest master of Haystack
 !pip install git+https://github.com/deepset-ai/haystack.git
 !pip install urllib3==1.25.4
-!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html
-
 ```
 
 
@@ -136,7 +134,7 @@ print(df.head())
 # Get embeddings for our questions from the FAQs
 questions = list(df["question"].values)
 df["question_emb"] = retriever.embed_queries(texts=questions)
-df = df.rename(columns={"answer": "text"})
+df = df.rename(columns={"question": "text"})
 
 # Convert Dataframe to list of dicts and index them in our DocumentStore
 docs_to_index = df.to_dict(orient="records")
@@ -144,11 +142,18 @@ document_store.write_documents(docs_to_index)
 ```
 
 ### Ask questions
-Initialize a Finder (this time without a reader) and ask questions
+Initialize a Pipeline (this time without a reader) and ask questions
 
 
 ```python
-finder = Finder(reader=None, retriever=retriever)
+from haystack.pipeline import FAQPipeline
-prediction = finder.get_answers_via_similar_questions(question="How is the virus spreading?", top_k_retriever=10)
+pipe = FAQPipeline(retriever=retriever)
-print_answers(prediction, details="all")
+```
+
+
+```python
+prediction = pipe.run(query="How is the virus spreading?", top_k_retriever=10)
+print_answers(prediction, details="all")
+
+
 ```

docs/_src/tutorials/tutorials/5.md

@@ -38,8 +38,6 @@ You can start Elasticsearch on your local machine instance using Docker. If Dock
 # Install the latest master of Haystack
 !pip install git+https://github.com/deepset-ai/haystack.git
 !pip install urllib3==1.25.4
-!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html
-
 ```
 
 
@@ -177,7 +175,6 @@ print("Reader F1-Score:", reader_eval_results["f1"])
 
 ```python
 # Evaluate combination of Reader and Retriever through Finder
-# Evaluate combination of Reader and Retriever through Finder
 finder_eval_results = finder.eval(top_k_retriever=1, top_k_reader=10, label_index=label_index, doc_index=doc_index)
 finder.print_eval_results(finder_eval_results)
 ```

docs/_src/tutorials/tutorials/6.md

@@ -80,8 +80,6 @@ Make sure you enable the GPU runtime to experience decent speed in this tutorial
 # Install the latest master of Haystack
 !pip install git+https://github.com/deepset-ai/haystack.git
 !pip install urllib3==1.25.4
-!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html
-
 ```
 
 
@@ -179,33 +177,31 @@ Here we use a FARMReader with the *deepset/roberta-base-squad2* model (see: http
 reader = FARMReader(model_name_or_path="deepset/roberta-base-squad2", use_gpu=True)
 ```
 
-#### Finder
+### Pipeline
 
-The Finder sticks together reader and retriever in a pipeline to answer our actual questions. 
+With a Haystack `Pipeline` you can stick together your building blocks to a search pipeline.
+Under the hood, `Pipelines` are Directed Acyclic Graphs (DAGs) that you can easily customize for your own use cases.
+To speed things up, Haystack also comes with a few predefined Pipelines. One of them is the `ExtractiveQAPipeline` that combines a retriever and a reader to answer our questions.
+You can learn more about `Pipelines` in the [docs](https://haystack.deepset.ai/docs/latest/pipelinesmd).
 
 
 ```python
-finder = Finder(reader, retriever)
+from haystack.pipeline import ExtractiveQAPipeline
+pipe = ExtractiveQAPipeline(reader, retriever)
 ```
 
-### Voilà! Ask a question!
+## Voilà! Ask a question!
 
 
 ```python
 # You can configure how many candidates the reader and retriever shall return
-# The higher top_k_retriever, the better (but also the slower) your answers. 
+# The higher top_k_retriever, the better (but also the slower) your answers.
-prediction = finder.get_answers(question="Who created the Dothraki vocabulary?", top_k_retriever=10, top_k_reader=5)
+prediction = pipe.run(query="Who created the Dothraki vocabulary?", top_k_retriever=10, top_k_reader=5)
-
-#prediction = finder.get_answers(question="Who is the father of Arya Stark?", top_k_retriever=10, top_k_reader=5)
-#prediction = finder.get_answers(question="Who is the sister of Sansa?", top_k_retriever=10, top_k_reader=5)
 ```
 
 
 ```python
 print_answers(prediction, details="minimal")
-```
 
 
-```python
-
 ```

docs/_src/tutorials/tutorials/7.md

@@ -37,8 +37,6 @@ Here are the packages and imports that we'll need:
 ```python
 !pip install git+https://github.com/deepset-ai/haystack.git
 !pip install urllib3==1.25.4
-!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html
-
 ```
 
 
@@ -188,3 +186,14 @@ for question in QUESTIONS:
     answers = predicted_result["answers"]
     print(f'Generated answer is \'{answers[0]["answer"]}\' for the question = \'{question}\'')
 ```
+
+
+```python
+# Or alternatively use the Pipeline class
+from haystack.pipeline import GenerativeQAPipeline
+
+pipe = GenerativeQAPipeline(generator=generator, retriever=retriever)
+for question in QUESTIONS:
+    res = pipe.run(query=question, top_k_generator=1, top_k_retriever=5)
+    print(res)
+```

docs/_src/tutorials/tutorials/8.md

@@ -36,9 +36,8 @@ This tutorial will show you all the tools that Haystack provides to help you cas
 
 # Install the latest master of Haystack
 !pip install git+https://github.com/deepset-ai/haystack.git
-!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html
+!wget --no-check-certificate https://dl.xpdfreader.com/xpdf-tools-linux-4.03.tar.gz
-!wget --no-check-certificate https://dl.xpdfreader.com/xpdf-tools-linux-4.02.tar.gz
+!tar -xvf xpdf-tools-linux-4.03.tar.gz && sudo cp xpdf-tools-linux-4.03/bin64/pdftotext /usr/local/bin
-!tar -xvf xpdf-tools-linux-4.02.tar.gz && sudo cp xpdf-tools-linux-4.02/bin64/pdftotext /usr/local/bin
 ```
 
 

docs/_src/tutorials/tutorials/9.md

@@ -21,7 +21,6 @@ This tutorial will guide you through the steps required to create a retriever th
 
 # Install the latest master of Haystack
 !pip install git+https://github.com/deepset-ai/haystack.git
-!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html
 ```
 
 

haystack/pipeline.py

@@ -434,7 +434,7 @@ class FAQPipeline(BaseStandardPipeline):
 
         results: Dict = {"query": query, "answers": []}
         for doc in documents:
-            # TODO proper calibratation of pseudo probabilities
+            # TODO proper calibration of pseudo probabilities
             cur_answer = {
                 "query": doc.text,
                 "answer": doc.meta["answer"],
@@ -448,7 +448,6 @@ class FAQPipeline(BaseStandardPipeline):
             }
 
             results["answers"].append(cur_answer)
-
         return results
 
 

haystack/retriever/squad_to_dpr.py

@@ -98,7 +98,7 @@ class HaystackDocumentStore:
         if not es.ping():
             logging.info("Starting Elasticsearch ...")
             status = subprocess.run(
-                ['docker run -d -p 9200:9200 -e "discovery.type=single-node" elasticsearch:7.6.2'], shell=True
+                ['docker run -d -p 9200:9200 -e "discovery.type=single-node" elasticsearch:7.9.2'], shell=True
             )
             if status.returncode:
                 raise Exception(

run_docker_gpu.sh

@@ -6,7 +6,7 @@
 #
 # To use GPU with Docker, ensure nvidia-docker(https://github.com/NVIDIA/nvidia-docker) is installed.
 
-docker run -d -p 9200:9200 -e "discovery.type=single-node" elasticsearch:7.6.1
+docker run -d -p 9200:9200 -e "discovery.type=single-node" elasticsearch:7.9.2
 # alternative: for a demo you can also use this elasticsearch image with already indexed GoT articles
 #docker run -d -p 9200:9200 -e "discovery.type=single-node" deepset/elasticsearch-game-of-thrones
 

tutorials/Tutorial1_Basic_QA_Pipeline.ipynb

@@ -60,8 +60,7 @@
     "\n",
     "# Install the latest master of Haystack\n",
     "!pip install git+https://github.com/deepset-ai/haystack.git\n",
-    "!pip install urllib3==1.25.4\n",
+    "!pip install urllib3==1.25.4\n"
-    "!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html\n"
    ]
   },
   {
@@ -70,7 +69,6 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "from haystack import Finder\n",
     "from haystack.preprocessor.cleaning import clean_wiki_text\n",
     "from haystack.preprocessor.utils import convert_files_to_dicts, fetch_archive_from_http\n",
     "from haystack.reader.farm import FARMReader\n",
@@ -111,7 +109,7 @@
    ],
    "source": [
     "# Recommended: Start Elasticsearch using Docker\n",
-    "#! docker run -d -p 9200:9200 -e \"discovery.type=single-node\" elasticsearch:7.6.2"
+    "#! docker run -d -p 9200:9200 -e \"discovery.type=single-node\" elasticsearch:7.9.2"
    ]
   },
   {
@@ -121,13 +119,13 @@
    "outputs": [],
    "source": [
     "# In Colab / No Docker environments: Start Elasticsearch from source\n",
-    "! wget https://artifacts.elastic.co/downloads/elasticsearch/elasticsearch-7.6.2-linux-x86_64.tar.gz -q\n",
+    "! wget https://artifacts.elastic.co/downloads/elasticsearch/elasticsearch-7.9.2-linux-x86_64.tar.gz -q\n",
-    "! tar -xzf elasticsearch-7.6.2-linux-x86_64.tar.gz\n",
+    "! tar -xzf elasticsearch-7.9.2-linux-x86_64.tar.gz\n",
-    "! chown -R daemon:daemon elasticsearch-7.6.2\n",
+    "! chown -R daemon:daemon elasticsearch-7.9.2\n",
     "\n",
     "import os\n",
     "from subprocess import Popen, PIPE, STDOUT\n",
-    "es_server = Popen(['elasticsearch-7.6.2/bin/elasticsearch'],\n",
+    "es_server = Popen(['elasticsearch-7.9.2/bin/elasticsearch'],\n",
     "                   stdout=PIPE, stderr=STDOUT,\n",
     "                   preexec_fn=lambda: os.setuid(1)  # as daemon\n",
     "                  )\n",
@@ -359,9 +357,12 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "### Finder\n",
+    "### Pipeline\n",
     "\n",
-    "The Finder sticks together reader and retriever in a pipeline to answer our actual questions. "
+    "With a Haystack `Pipeline` you can stick together your building blocks to a search pipeline.\n",
+    "Under the hood, `Pipelines` are Directed Acyclic Graphs (DAGs) that you can easily customize for your own use cases.\n",
+    "To speed things up, Haystack also comes with a few predefined Pipelines. One of them is the `ExtractiveQAPipeline` that combines a retriever and a reader to answer our questions.\n",
+    "You can learn more about `Pipelines` in the [docs](https://haystack.deepset.ai/docs/latest/pipelinesmd)."
    ]
   },
   {
@@ -374,7 +375,8 @@
    },
    "outputs": [],
    "source": [
-    "finder = Finder(reader, retriever)"
+    "from haystack.pipeline import ExtractiveQAPipeline\n",
+    "pipe = ExtractiveQAPipeline(reader, retriever)"
    ]
   },
   {
@@ -406,7 +408,7 @@
    "source": [
     "# You can configure how many candidates the reader and retriever shall return\n",
     "# The higher top_k_retriever, the better (but also the slower) your answers. \n",
-    "prediction = finder.get_answers(question=\"Who is the father of Arya Stark?\", top_k_retriever=10, top_k_reader=5)"
+    "prediction = pipe.run(query=\"Who is the father of Arya Stark?\", top_k_retriever=10, top_k_reader=5)"
    ]
   },
   {
@@ -415,8 +417,8 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# prediction = finder.get_answers(question=\"Who created the Dothraki vocabulary?\", top_k_reader=5)\n",
+    "# prediction = pipe.run(query=\"Who created the Dothraki vocabulary?\", top_k_reader=5)\n",
-    "# prediction = finder.get_answers(question=\"Who is the sister of Sansa?\", top_k_reader=5)"
+    "# prediction = pipe.run(query=\"Who is the sister of Sansa?\", top_k_reader=5)"
    ]
   },
   {

tutorials/Tutorial1_Basic_QA_Pipeline.py

@@ -48,7 +48,7 @@ def tutorial1_basic_qa_pipeline():
     if LAUNCH_ELASTICSEARCH:
         logging.info("Starting Elasticsearch ...")
         status = subprocess.run(
-            ['docker run -d -p 9200:9200 -e "discovery.type=single-node" elasticsearch:7.6.2'], shell=True
+            ['docker run -d -p 9200:9200 -e "discovery.type=single-node" elasticsearch:7.9.2'], shell=True
         )
         if status.returncode:
             raise Exception("Failed to launch Elasticsearch. If you want to connect to an existing Elasticsearch instance"
@@ -138,20 +138,20 @@ def tutorial1_basic_qa_pipeline():
     # reader = TransformersReader(
     #    model_name_or_path="distilbert-base-uncased-distilled-squad", tokenizer="distilbert-base-uncased", use_gpu=-1)
 
-    # ### Finder
+    # ### Pipeline
-    #
+    # 
-    # The Finder sticks together reader and retriever in a pipeline to answer our actual questions.
+    # With a Haystack `Pipeline` you can stick together your building blocks to a search pipeline.
+    # Under the hood, `Pipelines` are Directed Acyclic Graphs (DAGs) that you can easily customize for your own use cases.
+    # To speed things up, Haystack also comes with a few predefined Pipelines. One of them is the `ExtractiveQAPipeline` that combines a retriever and a reader to answer our questions.
+    # You can learn more about `Pipelines` in the [docs](https://haystack.deepset.ai/docs/latest/pipelinesmd).
+    from haystack.pipeline import ExtractiveQAPipeline
+    pipe = ExtractiveQAPipeline(reader, retriever)
+    
+    ## Voilà! Ask a question!
+    prediction = pipe.run(query="Who is the father of Arya Stark?", top_k_retriever=10, top_k_reader=5)
 
-    finder = Finder(reader, retriever)
+    # prediction = pipe.run(query="Who created the Dothraki vocabulary?", top_k_reader=5)
-
+    # prediction = pipe.run(query="Who is the sister of Sansa?", top_k_reader=5)
-    # ## Voilà! Ask a question!
-    # You can configure how many candidates the reader and retriever shall return
-    # The higher top_k_retriever, the better (but also the slower) your answers.
-    prediction = finder.get_answers(question="Who is the father of Sansa Stark?", top_k_retriever=10, top_k_reader=5)
-
-
-    # prediction = finder.get_answers(question="Who created the Dothraki vocabulary?", top_k_reader=5)
-    # prediction = finder.get_answers(question="Who is the sister of Sansa?", top_k_reader=5)
 
     print_answers(prediction, details="minimal")
 

tutorials/Tutorial2_Finetune_a_model_on_your_data.ipynb

@@ -56,8 +56,7 @@
     "\n",
     "# Install the latest master of Haystack\n",
     "!pip install git+https://github.com/deepset-ai/haystack.git\n",
-    "!pip install urllib3==1.25.4\n",
+    "!pip install urllib3==1.25.4"
-    "!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html\n"
    ]
   },
   {

tutorials/Tutorial3_Basic_QA_Pipeline_without_Elasticsearch.ipynb

@@ -56,8 +56,7 @@
     "\n",
     "# Install the latest master of Haystack\n",
     "!pip install git+https://github.com/deepset-ai/haystack.git\n",
-    "!pip install urllib3==1.25.4\n",
+    "!pip install urllib3==1.25.4"
-    "!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html\n"
    ]
   },
   {
@@ -278,9 +277,12 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "### Finder\n",
+    "### Pipeline\n",
     "\n",
-    "The Finder sticks together reader and retriever in a pipeline to answer our actual questions. "
+    "With a Haystack `Pipeline` you can stick together your building blocks to a search pipeline.\n",
+    "Under the hood, `Pipelines` are Directed Acyclic Graphs (DAGs) that you can easily customize for your own use cases.\n",
+    "To speed things up, Haystack also comes with a few predefined Pipelines. One of them is the `ExtractiveQAPipeline` that combines a retriever and a reader to answer our questions.\n",
+    "You can learn more about `Pipelines` in the [docs](https://haystack.deepset.ai/docs/latest/pipelinesmd)."
    ]
   },
   {
@@ -293,7 +295,8 @@
    },
    "outputs": [],
    "source": [
-    "finder = Finder(reader, retriever)"
+    "from haystack.pipeline import ExtractiveQAPipeline\n",
+    "pipe = ExtractiveQAPipeline(reader, retriever)"
    ]
   },
   {
@@ -340,8 +343,8 @@
    ],
    "source": [
     "# You can configure how many candidates the reader and retriever shall return\n",
-    "# The higher top_k_retriever, the better (but also the slower) your answers. \n",
+    "# The higher top_k_retriever, the better (but also the slower) your answers.\n",
-    "prediction = finder.get_answers(question=\"Who is the father of Arya Stark?\", top_k_retriever=10, top_k_reader=5)"
+    "prediction = pipe.run(query=\"Who is the father of Arya Stark?\", top_k_retriever=10, top_k_reader=5)"
    ]
   },
   {
@@ -350,8 +353,8 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# prediction = finder.get_answers(question=\"Who created the Dothraki vocabulary?\", top_k_reader=5)\n",
+    "# prediction = pipe.run(query=\"Who created the Dothraki vocabulary?\", top_k_reader=5)\n",
-    "# prediction = finder.get_answers(question=\"Who is the sister of Sansa?\", top_k_reader=5)"
+    "# prediction = pipe.run(query=\"Who is the sister of Sansa?\", top_k_reader=5)"
    ]
   },
   {

tutorials/Tutorial3_Basic_QA_Pipeline_without_Elasticsearch.py

@@ -91,18 +91,20 @@ def tutorial3_basic_qa_pipeline_without_elasticsearch():
     # Alternative:
     # reader = TransformersReader(model_name_or_path="distilbert-base-uncased-distilled-squad", tokenizer="distilbert-base-uncased", use_gpu=-1)
 
-    # ### Finder
+    # ### Pipeline
     #
-    # The Finder sticks together reader and retriever in a pipeline to answer our actual questions.
+    # With a Haystack `Pipeline` you can stick together your building blocks to a search pipeline.
-    finder = Finder(reader, retriever)
+    # Under the hood, `Pipelines` are Directed Acyclic Graphs (DAGs) that you can easily customize for your own use cases.
+    # To speed things up, Haystack also comes with a few predefined Pipelines. One of them is the `ExtractiveQAPipeline` that combines a retriever and a reader to answer our questions.
+    # You can learn more about `Pipelines` in the [docs](https://haystack.deepset.ai/docs/latest/pipelinesmd).
+    from haystack.pipeline import ExtractiveQAPipeline
+    pipe = ExtractiveQAPipeline(reader, retriever)
 
-    # ## Voilà! Ask a question!
+    ## Voilà! Ask a question!
+    prediction = pipe.run(query="Who is the father of Arya Stark?", top_k_retriever=10, top_k_reader=5)
 
-    # You can configure how many candidates the reader and retriever shall return
+    # prediction = pipe.run(query="Who created the Dothraki vocabulary?", top_k_reader=5)
-    # The higher top_k_retriever, the better (but also the slower) your answers.
+    # prediction = pipe.run(query="Who is the sister of Sansa?", top_k_reader=5)
-    prediction = finder.get_answers(question="Who is the father of Arya Stark?", top_k_retriever=10, top_k_reader=5)
-    # prediction = finder.get_answers(question="Who created the Dothraki vocabulary?", top_k_reader=5)
-    # prediction = finder.get_answers(question="Who is the sister of Sansa?", top_k_reader=5)
 
     print_answers(prediction, details="minimal")
 

tutorials/Tutorial4_FAQ_style_QA.ipynb

@@ -64,8 +64,7 @@
     "\n",
     "# Install the latest master of Haystack\n",
     "!pip install git+https://github.com/deepset-ai/haystack.git\n",
-    "!pip install urllib3==1.25.4\n",
+    "!pip install urllib3==1.25.4"
-    "!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html\n"
    ]
   },
   {
@@ -221,7 +220,7 @@
     "# Get embeddings for our questions from the FAQs\n",
     "questions = list(df[\"question\"].values)\n",
     "df[\"question_emb\"] = retriever.embed_queries(texts=questions)\n",
-    "df = df.rename(columns={\"answer\": \"text\"})\n",
+    "df = df.rename(columns={\"question\": \"text\"})\n",
     "\n",
     "# Convert Dataframe to list of dicts and index them in our DocumentStore\n",
     "docs_to_index = df.to_dict(orient=\"records\")\n",
@@ -238,7 +237,7 @@
    "cell_type": "markdown",
    "source": [
     "### Ask questions\n",
-    "Initialize a Finder (this time without a reader) and ask questions"
+    "Initialize a Pipeline (this time without a reader) and ask questions"
    ],
    "metadata": {
     "collapsed": false
@@ -249,9 +248,8 @@
    "execution_count": null,
    "outputs": [],
    "source": [
-    "finder = Finder(reader=None, retriever=retriever)\n",
+    "from haystack.pipeline import FAQPipeline\n",
-    "prediction = finder.get_answers_via_similar_questions(question=\"How is the virus spreading?\", top_k_retriever=10)\n",
+    "pipe = FAQPipeline(retriever=retriever)"
-    "print_answers(prediction, details=\"all\")"
    ],
    "metadata": {
     "collapsed": false,
@@ -259,6 +257,22 @@
      "name": "#%%\n"
     }
    }
+  },
+  {
+   "cell_type": "code",
+   "source": [
+    "prediction = pipe.run(query=\"How is the virus spreading?\", top_k_retriever=10)\n",
+    "print_answers(prediction, details=\"all\")\n",
+    "\n"
+   ],
+   "metadata": {
+    "collapsed": false,
+    "pycharm": {
+     "name": "#%%\n"
+    }
+   },
+   "execution_count": null,
+   "outputs": []
   }
  ],
  "metadata": {

tutorials/Tutorial4_FAQ_style_QA.py

@@ -24,7 +24,7 @@ def tutorial4_faq_style_qa():
     # - Generalizability: We can only answer questions that are similar to existing ones in FAQ
     #
     # In some use cases, a combination of extractive QA and FAQ-style can also be an interesting option.
-    LAUNCH_ELASTICSEARCH=True
+    LAUNCH_ELASTICSEARCH=False
 
     if LAUNCH_ELASTICSEARCH:
         logging.info("Starting Elasticsearch ...")
@@ -34,7 +34,7 @@ def tutorial4_faq_style_qa():
         if status.returncode:
             raise Exception("Failed to launch Elasticsearch. If you want to connect to an existing Elasticsearch instance"
                             "then set LAUNCH_ELASTICSEARCH in the script to False.")
-        time.sleep(15)
+        time.sleep(30)
 
     ### Init the DocumentStore
     # In contrast to Tutorial 1 (extractive QA), we:
@@ -71,16 +71,18 @@ def tutorial4_faq_style_qa():
     # Get embeddings for our questions from the FAQs
     questions = list(df["question"].values)
     df["question_emb"] = retriever.embed_queries(texts=questions)
-    df = df.rename(columns={"answer": "text"})
+    df = df.rename(columns={"question": "text"})
 
     # Convert Dataframe to list of dicts and index them in our DocumentStore
     docs_to_index = df.to_dict(orient="records")
     document_store.write_documents(docs_to_index)
 
+    #    Initialize a Pipeline (this time without a reader) and ask questions
 
-    # Init reader & and use Finder to get answer (same as in Tutorial 1)
+    from haystack.pipeline import FAQPipeline
-    finder = Finder(reader=None, retriever=retriever)
+    pipe = FAQPipeline(retriever=retriever)
-    prediction = finder.get_answers_via_similar_questions(question="How is the virus spreading?", top_k_retriever=10)
+
+    prediction = pipe.run(query="How is the virus spreading?", top_k_retriever=10)
     print_answers(prediction, details="all")
 
 

tutorials/Tutorial5_Evaluation.ipynb

@@ -77,8 +77,7 @@
     "\n",
     "# Install the latest master of Haystack\n",
     "!pip install git+https://github.com/deepset-ai/haystack.git\n",
-    "!pip install urllib3==1.25.4\n",
+    "!pip install urllib3==1.25.4"
-    "!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html\n"
    ]
   },
   {
@@ -469,7 +468,6 @@
    },
    "outputs": [],
    "source": [
-    "# Evaluate combination of Reader and Retriever through Finder\n",
     "# Evaluate combination of Reader and Retriever through Finder\n",
     "finder_eval_results = finder.eval(top_k_retriever=1, top_k_reader=10, label_index=label_index, doc_index=doc_index)\n",
     "finder.print_eval_results(finder_eval_results)"

tutorials/Tutorial6_Better_Retrieval_via_DPR.ipynb

@@ -286,8 +286,7 @@
     "\n",
     "# Install the latest master of Haystack\n",
     "!pip install git+https://github.com/deepset-ai/haystack.git\n",
-    "!pip install urllib3==1.25.4\n",
+    "!pip install urllib3==1.25.4"
-    "!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html\n"
    ]
   },
   {
@@ -647,9 +646,12 @@
     "id": "unhLD18yA6OF"
    },
    "source": [
-    "#### Finder\n",
+    "### Pipeline\n",
     "\n",
-    "The Finder sticks together reader and retriever in a pipeline to answer our actual questions. "
+    "With a Haystack `Pipeline` you can stick together your building blocks to a search pipeline.\n",
+    "Under the hood, `Pipelines` are Directed Acyclic Graphs (DAGs) that you can easily customize for your own use cases.\n",
+    "To speed things up, Haystack also comes with a few predefined Pipelines. One of them is the `ExtractiveQAPipeline` that combines a retriever and a reader to answer our questions.\n",
+    "You can learn more about `Pipelines` in the [docs](https://haystack.deepset.ai/docs/latest/pipelinesmd)."
    ]
   },
   {
@@ -662,7 +664,8 @@
    },
    "outputs": [],
    "source": [
-    "finder = Finder(reader, retriever)"
+    "from haystack.pipeline import ExtractiveQAPipeline\n",
+    "pipe = ExtractiveQAPipeline(reader, retriever)"
    ]
   },
   {
@@ -672,7 +675,7 @@
     "id": "bXlBBxKXA6OL"
    },
    "source": [
-    "### Voilà! Ask a question!"
+    "## Voilà! Ask a question!"
    ]
   },
   {
@@ -712,65 +715,8 @@
    ],
    "source": [
     "# You can configure how many candidates the reader and retriever shall return\n",
-    "# The higher top_k_retriever, the better (but also the slower) your answers. \n",
+    "# The higher top_k_retriever, the better (but also the slower) your answers.\n",
-    "prediction = finder.get_answers(question=\"Who created the Dothraki vocabulary?\", top_k_retriever=10, top_k_reader=5)\n",
+    "prediction = pipe.run(query=\"Who created the Dothraki vocabulary?\", top_k_retriever=10, top_k_reader=5)"
-    "\n",
-    "#prediction = finder.get_answers(question=\"Who is the father of Arya Stark?\", top_k_retriever=10, top_k_reader=5)\n",
-    "#prediction = finder.get_answers(question=\"Who is the sister of Sansa?\", top_k_retriever=10, top_k_reader=5)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 9,
-   "metadata": {
-    "colab": {
-     "base_uri": "https://localhost:8080/",
-     "height": 561
-    },
-    "colab_type": "code",
-    "id": "N70FgfkwA6OQ",
-    "outputId": "9419c75d-181c-4ef6-cea8-b328a503f19a",
-    "pycharm": {
-     "name": "#%%\n"
-    }
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "[   {   'answer': 'David J. Peterson',\n",
-      "        'context': '\\n'\n",
-      "                   '===Valyrian===\\n'\n",
-      "                   'David J. Peterson, who created the Dothraki language for '\n",
-      "                   'the first season of the show, was entrusted by the '\n",
-      "                   'producers to design a new '},\n",
-      "    {   'answer': 'David Peterson',\n",
-      "        'context': '\\n'\n",
-      "                   '==Phonology and romanization==\\n'\n",
-      "                   'David Peterson has said, \"You know, most people probably '\n",
-      "                   \"don't really know what Arabic actually sounds like, so to \"\n",
-      "                   'an '},\n",
-      "    {   'answer': 'books',\n",
-      "        'context': 'ints.  First, the language had to match the uses already '\n",
-      "                   'put down in the books. Secondly, it had to be easily '\n",
-      "                   'pronounceable or learnable by the actors'},\n",
-      "    {   'answer': \"'''Nevakhi vekha ha maan: Rekke, m'aresakea norethi fitte.'\",\n",
-      "        'context': '\\n'\n",
-      "                   '==Sample==\\n'\n",
-      "                   \": '''Nevakhi vekha ha maan: Rekke, m'aresakea norethi \"\n",
-      "                   \"fitte.'''\\n\"\n",
-      "                   ': seat. exist. for  there. with.coward. hair. short\\n'\n",
-      "                   \": ''There is a place f\"},\n",
-      "    {   'answer': 'Tyrion',\n",
-      "        'context': 'ding, as well as his nephew Joffrey, the new king, as '\n",
-      "                   'civil war begins. Tyrion struggles to strengthen and '\n",
-      "                   'protect the city and family who hate him an'}]\n"
-     ]
-    }
-   ],
-   "source": [
-    "print_answers(prediction, details=\"minimal\")"
    ]
   },
   {
@@ -778,7 +724,10 @@
    "execution_count": null,
    "metadata": {},
    "outputs": [],
-   "source": []
+   "source": [
+    "print_answers(prediction, details=\"minimal\")\n",
+    "\n"
+   ]
   }
  ],
  "metadata": {

tutorials/Tutorial6_Better_Retrieval_via_DPR.py

@@ -52,18 +52,15 @@ def tutorial6_better_retrieval_via_dpr():
     # Hugging Face's model hub (https://huggingface.co/models)
     reader = FARMReader(model_name_or_path="deepset/roberta-base-squad2", use_gpu=True)
 
-    ### Finder
+    ### Pipeline
-    # The Finder sticks together reader and retriever in a pipeline to answer our actual questions.
+    from haystack.pipeline import ExtractiveQAPipeline
-    finder = Finder(reader, retriever)
+    pipe = ExtractiveQAPipeline(reader, retriever)
 
-    ### Voilà! Ask a question!
+    ## Voilà! Ask a question!
-    # You can configure how many candidates the reader and retriever shall return
+    prediction = pipe.run(query="Who is the father of Arya Stark?", top_k_retriever=10, top_k_reader=5)
-    # The higher top_k_retriever, the better (but also the slower) your answers.
-    prediction = finder.get_answers(question="Who is the father of Arya Stark?", top_k_retriever=10, top_k_reader=5)
 
-
+    # prediction = pipe.run(query="Who created the Dothraki vocabulary?", top_k_reader=5)
-    # prediction = finder.get_answers(question="Who created the Dothraki vocabulary?", top_k_reader=5)
+    # prediction = pipe.run(query="Who is the sister of Sansa?", top_k_reader=5)
-    # prediction = finder.get_answers(question="Who is the sister of Sansa?", top_k_reader=5)
 
     print_answers(prediction, details="minimal")
 

tutorials/Tutorial7_RAG_Generator.ipynb

@@ -62,8 +62,7 @@
    "outputs": [],
    "source": [
     "!pip install git+https://github.com/deepset-ai/haystack.git\n",
-    "!pip install urllib3==1.25.4\n",
+    "!pip install urllib3==1.25.4"
-    "!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html\n"
    ],
    "metadata": {
     "collapsed": false,
@@ -322,6 +321,26 @@
      "name": "#%%\n"
     }
    }
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": [],
+   "source": [
+    "# Or alternatively use the Pipeline class\n",
+    "from haystack.pipeline import GenerativeQAPipeline\n",
+    "\n",
+    "pipe = GenerativeQAPipeline(generator=generator, retriever=retriever)\n",
+    "for question in QUESTIONS:\n",
+    "    res = pipe.run(query=question, top_k_generator=1, top_k_retriever=5)\n",
+    "    print(res)"
+   ],
+   "metadata": {
+    "collapsed": false,
+    "pycharm": {
+     "name": "#%%\n"
+    }
+   }
   }
  ],
  "metadata": {

tutorials/Tutorial7_RAG_Generator.py

@@ -111,6 +111,12 @@ def tutorial7_rag_generator():
         answers = predicted_result["answers"]
         print(f'Generated answer is \'{answers[0]["answer"]}\' for the question = \'{question}\'')
 
+    # Or alternatively use the Pipeline class
+    from haystack.pipeline import GenerativeQAPipeline
+    pipe = GenerativeQAPipeline(generator=generator, retriever=retriever)
+    for question in QUESTIONS:
+        res = pipe.run(query=question, top_k_generator=1, top_k_retriever=5)
+        print(res)
 
 if __name__ == "__main__":
     tutorial7_rag_generator()

tutorials/Tutorial8_Preprocessing.ipynb

@@ -59,9 +59,8 @@
     "\n",
     "# Install the latest master of Haystack\n",
     "!pip install git+https://github.com/deepset-ai/haystack.git\n",
-    "!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html\n",
+    "!wget --no-check-certificate https://dl.xpdfreader.com/xpdf-tools-linux-4.03.tar.gz\n",
-    "!wget --no-check-certificate https://dl.xpdfreader.com/xpdf-tools-linux-4.02.tar.gz\n",
+    "!tar -xvf xpdf-tools-linux-4.03.tar.gz && sudo cp xpdf-tools-linux-4.03/bin64/pdftotext /usr/local/bin"
-    "!tar -xvf xpdf-tools-linux-4.02.tar.gz && sudo cp xpdf-tools-linux-4.02/bin64/pdftotext /usr/local/bin"
    ],
    "metadata": {
     "collapsed": false,
@@ -497,15 +496,6 @@
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython2",
    "version": "2.7.6"
-  },
-  "pycharm": {
-   "stem_cell": {
-    "cell_type": "raw",
-    "source": [],
-    "metadata": {
-     "collapsed": false
-    }
-   }
   }
  },
  "nbformat": 4,

tutorials/Tutorial9_DPR_training.ipynb

@@ -26,8 +26,7 @@
     "#! pip install farm-haystack\n",
     "\n",
     "# Install the latest master of Haystack\n",
-    "!pip install git+https://github.com/deepset-ai/haystack.git\n",
+    "!pip install git+https://github.com/deepset-ai/haystack.git"
-    "!pip install torch==1.6.0+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html"
    ],
    "metadata": {
     "collapsed": false,
@@ -403,15 +402,6 @@
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython2",
    "version": "2.7.6"
-  },
-  "pycharm": {
-   "stem_cell": {
-    "cell_type": "raw",
-    "source": [],
-    "metadata": {
-     "collapsed": false
-    }
-   }
   }
  },
  "nbformat": 4,