added code for ner task

Author: danielecalda

.gitignore

@@ -0,0 +1,4 @@
+venv
+.ipynb_checkpoints
+models/
+.idea

README.md

@@ -1,2 +1,51 @@
-# blog-nlp
-This is repository for nlp tasks related nlp articles on tech blog
+# Blog NLP
+
+In this repository are present some jupyter notebooks and scripts related about Tech Blog Natural
+Language Processing articles series.
+
+## Repository Organization
+
+This repository is organized in multiple folders. Every folder is related to specific article in Smc Tech Blog.
+
+* [1-ner](1-ner): contains code related to Named Entity Recognition article
+* [2-text-classification](): coming soon
+* [3-topic-modeling](): coming soon
+* [4-entity-linking](): coming soon
+* [5-machine-translation](): coming soon
+
+## Execution in Docker Environment
+
+All code is executable in Docker environment. There is a Docker Compose file that build an image based
+on Tensorflow with Jupyter Notebook integrated. (See Dockerfile inside nlp folder)
+The image contains also some default notebooks provided by tensorflow for some simple tutorials.
+Inside nlp folder are presente the directiories for every task based on repository organization.
+
+To build and run the container execute:
+
+```shell
+docker-compose up -d
+```
+
+You can see logs of servive typing:
+
+```shell
+docker-compose logs -f nlp
+```
+
+
+To rebuild image type:
+
+```shell
+docker-compose build
+```
+
+To stop container run:
+
+```shell
+docker-compose down
+```
+
+The nlp directory is binded as external volumes, so all changes in notebooks inside container 
+are saved.
+
+

docker-compose.yml

@@ -0,0 +1,12 @@
+version: '3.5'
+services:
+    nlp:
+        container_name: nlp
+        build:
+            dockerfile: nlp/Dockerfile
+            context: .
+        volumes:
+            - ./nlp:/tf/nlp
+        ports:
+            - '8888:8888'
+            - '5000:5000'

nlp/1-ner/README.md

@@ -0,0 +1,12 @@
+# Blog NLP
+
+In this directory is present related to [Tech Blog Named Entity Recognition article]()
+
+## Repository Organization
+
+There are three folders:
+
+* [notebooks](notebooks): container jupyter notebooks to train ner model with Spacy and Tensoflow Keras
+* [data](data): contains data to train models in BILUO, IOB and JSONL format. Contains also notebooks with code
+to transform data from one format to another
+* [models](models): is used to save trained models