Merge pull request #36 from ovh-ux/feat_live_incident

Feat live incident

CHANGELOG.md

@@ -1,3 +1,46 @@
+<a name="1.2.0"></a>
+# :gem: [1.2.0](https://github.com/ovh-ux/ovh-chatbot/compare/1.1.0...1.2.0) inexpensive-science (2017-08-30) :gem:
+
+
+### :ambulance: Bug Fixes :ambulance:
+
+  **controller/web:**
+
+  * fix promise in response not resolved ([0ce998f](https://github.com/ovh-ux/ovh-chatbot/commit/0ce998f))
+
+  **platforms:**
+
+  * fix reaching message length limit ([a34d1aa](https://github.com/ovh-ux/ovh-chatbot/commit/a34d1aa))
+
+  **security:**
+
+  * fix x-hub-signature wasn't properly checked ([cf7ad8e](https://github.com/ovh-ux/ovh-chatbot/commit/cf7ad8e))
+
+### :sparkles: Features :sparkles:
+
+  **apiai:**
+
+  * support for multiple apiai ([0de6eac](https://github.com/ovh-ux/ovh-chatbot/commit/0de6eac))
+
+  **bot:**
+
+  * add when will my service expires ? ([a6d7f3c](https://github.com/ovh-ux/ovh-chatbot/commit/a6d7f3c))
+  * inform user when service doesn't work anymore ([0330462](https://github.com/ovh-ux/ovh-chatbot/commit/0330462))
+
+  **express:**
+
+  * add cache for web users ([fe57549](https://github.com/ovh-ux/ovh-chatbot/commit/fe57549))
+
+  **Grunt:**
+
+  * add grunt for compiling api.ai zip configs ([a1ee7f7](https://github.com/ovh-ux/ovh-chatbot/commit/a1ee7f7))
+  * add i18n feature ([e971185](https://github.com/ovh-ux/ovh-chatbot/commit/e971185))
+  * add loggin ([d8ba0cf](https://github.com/ovh-ux/ovh-chatbot/commit/d8ba0cf))
+
+  **tools:**
+
+  * add api ai update task ([c4c7b81](https://github.com/ovh-ux/ovh-chatbot/commit/c4c7b81))
+
 <a name="1.1.0"></a>
 # :gem: [1.1.0](https://github.com/ovh-ux/ovh-chatbot/compare/1.0.0...v1.1.0) Moving Smoke (2017-08-07) :gem:
 

README.md

@@ -44,11 +44,28 @@ The bot uses [api.ai](https://www.api.ai) as a language processor.
 > + **Intent**: what the user *ask* to the bot.
 > + **Response**: What the bot *respond* to an user.
 
-All the intents needing dynamic responses can be found [here](bots/README.md). However, the responses to intents which doesn't need a dynamic response are stored in [api.ai](https://www.api.ai) database.
+All the intents needing processed responses can be found [here](bots/).
+
+The intents which doesn't need processed responses (aka: answers to basic questions) are stored in the [api.ai](https://www.api.ai) database, have a look at the [api.ai folder](apiai/).
+
+This intents and response are translated in:
++ Dutch
++ English (US, CA, GB)
++ French
++ German
++ Italian
++ Portuguese
++ Spanish
 
 ### Diagnostics
-Currently, this bot is able to diagnose a website, a xdsl line or a phone line.
-More details [here](diagnostics/README.md)
+Currently, this bot is able to:
++ diagnose a website
++ diagnose a xDSL line
++ diagnose a land line
++ warn about the expiration of a service
++ warn when an incident happens
+
+More details [here](diagnostics/)
 
 
 ## Contributing
@@ -67,10 +84,9 @@ Have a look at the [Contributing section](CONTRIBUTING.md). If you have any ques
 + Refactor to typescript
 + Docker compose
 + Documentation
-+ Translations
-+ When will my service expire?
-+ Monitoring
-+ Alerting
++ Cloud diagnostics (vps)
++ Order followup
++ Improve conversation skill
 
 ## License
 See https://github.com/ovh-ux/ovh-chatbot/blob/master/LICENSE

apiai/README.md

@@ -5,6 +5,7 @@
 > All the files related to apiai agents
 
 - `translations/`: contains the translation files.
+- `archives/`: contains an archive of the chatbot api.ai setup. It is used to generate the other archives
 - `template.json`: template file where every number is replaced by its corresponding value from the translations files
 
 ### Warning
@@ -22,6 +23,19 @@ In order to have the good api.ai token used (ie: use the right language), you ha
 }
 ```
 
+## Training apiai
+
+Once in production, you'll find that sometimes, api.ai doesn't find the right answer. Here is a step by step procedure to add new intent in production when you work with a preproduction agent.
+
+1. add the **new intents** to **preproduction**
+2. In the **production** agent, do the training with the **knwon** intents
+3. once you are done, export the **production** and import it in the **preproduction** (**DO NOT RESTORE**, you'll lose your progress)
+4. Now you can export your **preproduction** (it should have the new intents + the updated intents from production).
+5. You can test you archive by importing it in **production** and see in the training tab, if the answers are good.
+6. Repeat this process until your are satisfied.
+7. Once your done, follow the step in the [`tools`](../tools) folder to translate the agent
+
+
 ## tools
 
 Have a look at [/tools](../tools)!