Add langchain_text_splitters

tools/langchain_text_splitters/.shed.yml

@@ -0,0 +1,17 @@
+name: langchain_text_splitters
+owner: bgruening
+description: Split text into chunks using LangChain text splitters.
+long_description: |
+  Split plain-text-like datasets into chunks for LLM and RAG workflows using
+  langchain-text-splitters. The tool currently supports recursive character, character,
+  sentence-based splitting with NLTK or spaCy, and tiktoken-based token splitting,
+  and reports per-chunk lengths and start offsets as both readable text and
+  structured JSON.
+type: unrestricted
+categories:
+  - Natural Language Processing
+remote_repository_url: https://github.com/bgruening/galaxytools/tree/master/tools/langchain_text_splitters
+homepage_url: https://github.com/langchain-ai/langchain/tree/master/libs/text-splitters
+auto_tool_repositories:
+  name_template: "{{ tool_id }}"
+  description_template: "Wrapper for: {{ tool_name }}"

tools/langchain_text_splitters/macros.xml

@@ -0,0 +1,153 @@
+<macros>
+    <token name="@TOOL_VERSION@">1.1.2</token>
+    <token name="@VERSION_SUFFIX@">0</token>
+    <token name="@PROFILE@">25.1</token>
+    <!-- Expects $tiktoken_options to be set to the section holding the tokenizer settings. -->
+    <token name="@TIKTOKEN_ARGS@"><![CDATA[
+        #if str($tiktoken_options.tokenizer.tokenizer_selection_method) == "model"
+            --model-name '$tiktoken_options.tokenizer.model_name'
+        #else
+            --encoding-name '$tiktoken_options.tokenizer.encoding_name'
+        #end if
+        --allowed-special '$tiktoken_options.allowed_special'
+    ]]></token>
+    <xml name="requirements">
+        <requirements>
+            <requirement type="package" version="@TOOL_VERSION@">langchain-text-splitters</requirement>
+            <requirement type="package" version="0.13.0">tiktoken</requirement>
+            <requirement type="package" version="3.10.1">nltk</requirement>
+            <requirement type="package" version="2026.07.01">nltk_data</requirement>
+            <requirement type="package" version="3.8.14">spacy</requirement>
+            <requirement type="package" version="3.8.0">spacy-model-en_core_web_sm</requirement>
+            <!-- python-gil required for nltk -->
+            <requirement type="package" version="3.12.13">python-gil</requirement>
+        </requirements>
+    </xml>
+    <xml name="creators">
+        <creator>
+            <person givenName="Ivo Christopher" familyName="Leist" email="leist@informatik.uni-freiburg.de" identifier="https://orcid.org/0000-0002-9772-443X"/>
+            <person givenName="Arash" familyName="Kadkhodaei" url="http://github.com/arash77" identifier="https://orcid.org/0009-0006-2228-8123"/>
+            <organization name="AI4SOCIAL+" url="https://cordis.europa.eu/project/id/101292886"/>
+        </creator>
+    </xml>
+    <xml name="strip_whitespace" token_checked="true">
+        <param name="strip_whitespace" type="boolean" checked="@CHECKED@" truevalue="--strip-whitespace" falsevalue="" label="Strip whitespace around chunks" help="Remove leading and trailing whitespace from each generated chunk. Whitespace within chunks is preserved."/>
+    </xml>
+    <xml name="separator_definition">
+        <conditional name="separator_definition">
+            <param name="sep_opt" type="select" label="Separator">
+                <option value="paragraph_break" selected="true">Paragraph break (\n\n)</option>
+                <option value="line_break">Line break (\n)</option>
+                <option value="tab">Tab (\t)</option>
+                <option value="space">Space</option>
+                <option value="ascii_full_stop">ASCII full stop (.)</option>
+                <option value="ascii_comma">ASCII comma (,)</option>
+                <option value="ascii_semicolon">ASCII semicolon (;)</option>
+                <option value="zero_width_space">Zero-width space</option>
+                <option value="fullwidth_comma">Fullwidth comma</option>
+                <option value="ideographic_comma">Ideographic comma</option>
+                <option value="fullwidth_full_stop">Fullwidth full stop</option>
+                <option value="ideographic_full_stop">Ideographic full stop</option>
+                <option value="custom">Custom separator</option>
+            </param>
+            <when value="paragraph_break"/>
+            <when value="line_break"/>
+            <when value="tab"/>
+            <when value="space"/>
+            <when value="ascii_full_stop"/>
+            <when value="ascii_comma"/>
+            <when value="ascii_semicolon"/>
+            <when value="zero_width_space"/>
+            <when value="fullwidth_comma"/>
+            <when value="ideographic_comma"/>
+            <when value="fullwidth_full_stop"/>
+            <when value="ideographic_full_stop"/>
+            <when value="custom">
+                <param name="value" type="text" label="Custom separator" help="Enter a custom separator or separator sequence with its respective escape e.g. \n, \t, or \u3002">
+                    <!-- Characters that are neither listed here nor mapped by Galaxy are silently replaced
+                         by an X, which would split the text at the wrong places. The validator below rejects
+                         them instead, so that the user is pointed at the escape syntax. -->
+                    <sanitizer>
+                        <valid initial="default">
+                            <add value="\"/>
+                            <add value="#"/>
+                            <add value="|"/>
+                            <add value=";"/>
+                            <add value="~"/>
+                            <add value="%"/>
+                            <add value="&amp;"/>
+                            <add value="$"/>
+                            <add value="@"/>
+                            <add value="&lt;"/>
+                            <add value="&gt;"/>
+                            <add value="&quot;"/>
+                            <add value="["/>
+                            <add value="]"/>
+                            <add value="{"/>
+                            <add value="}"/>
+                        </valid>
+                    </sanitizer>
+                    <validator type="empty_field" message="Enter a custom separator."/>
+                    <validator type="regex" message="This separator contains a character that Galaxy cannot pass on unchanged, for example a letter outside the ASCII range or an apostrophe. Write such a character as an escape instead, for example \u3002 for the ideographic full stop.">^[A-Za-z0-9 \-=_.()/+*^,:?!\\#|;~%&amp;$@&lt;&gt;&quot;\[\]{}]+\Z</validator>
+                </param>
+            </when>
+        </conditional>
+    </xml>
+    <xml name="tiktoken_options">
+        <section name="tiktoken_options" title="Tokenization options" help="tiktoken is used to determine the number of tokens in a chunk. The OpenAI encoding and model name can be specified to ensure the token count is accurate for the downstream LLM workflow.">
+            <conditional name="tokenizer">
+                <param name="tokenizer_selection_method" type="select" label="Select tokenizer by" help="Choose an encoding directly or select a model whose associated encoding should be determined by tiktoken.">
+                    <option value="encoding" selected="true">Encoding name</option>
+                    <option value="model">Model name</option>
+                </param>
+                <when value="encoding">
+                    <param name="encoding_name" type="select" label="Encoding">
+                        <option value="o200k_harmony">o200k_harmony</option>
+                        <option value="o200k_base">o200k_base</option>
+                        <option value="cl100k_base">cl100k_base</option>
+                        <option value="p50k_edit">p50k_edit</option>
+                        <option value="p50k_base">p50k_base</option>
+                        <option value="r50k_base">r50k_base</option>
+                        <option value="gpt2" selected="true">gpt2</option>
+                    </param>
+                </when>
+                <when value="model">
+                    <param name="model_name" type="select" label="Model">
+                        <option value="o4-mini">o4-mini (reasoning)</option>
+                        <option value="o3">o3 (reasoning)</option>
+                        <option value="o1">o1 (reasoning)</option>
+                        <option value="gpt-5" selected="true">gpt-5 (chat)</option>
+                        <option value="gpt-4.1">gpt-4.1 (chat)</option>
+                        <option value="gpt-4o">gpt-4o (chat)</option>
+                        <option value="gpt-4">gpt-4 (chat)</option>
+                        <option value="gpt-3.5-turbo">gpt-3.5-turbo (chat)</option>
+                        <option value="gpt-3.5">gpt-3.5 (chat)</option>
+                        <option value="gpt-35-turbo">gpt-35-turbo (chat)</option>
+                        <option value="davinci-002">davinci-002 (base)</option>
+                        <option value="babbage-002">babbage-002 (base)</option>
+                        <option value="text-embedding-ada-002">text-embedding-ada-002 (embedding)</option>
+                        <option value="text-embedding-3-small">text-embedding-3-small (embedding)</option>
+                        <option value="text-embedding-3-large">text-embedding-3-large (embedding)</option>
+                    </param>
+                </when>
+            </conditional>
+            <param name="allowed_special" type="select" label="Special token handling">
+                <option value="none" selected="true">Reject special token strings</option>
+                <option value="all">Allow all special token strings</option>
+            </param>
+        </section>
+    </xml>
+    <xml name="chunk_params" token_size_label="Target chunk size" token_size_help="Desired chunk size, might be exceeded when the text between two separators is already larger than the target." token_overlap_label="Chunk overlap" token_overlap_help="Especially useful for retrieval-augmented generation (RAG) workflows, where overlapping text can help preserve context across chunk boundaries.">
+        <param name="chunk_size" type="integer" value="4000" min="1" label="@SIZE_LABEL@" help="@SIZE_HELP@"/>
+        <param name="chunk_overlap" type="integer" value="0" min="0" label="@OVERLAP_LABEL@" help="@OVERLAP_HELP@"/>
+    </xml>
+    <!-- Macros for testing -->
+    <xml name="assert_chunks_tsv" token_min_lines="2">
+        <output name="chunks_tsv">
+            <assert_contents>
+                <has_n_columns n="3"/>
+                <has_n_lines min="@MIN_LINES@"/>
+            </assert_contents>
+        </output>
+    </xml>
+</macros>